Abstract      

Agitation, such as verbal outbursts and physical aggression, is one of the most challenging symptoms of dementia in patients with Alzheimer’s disease. Non-pharmacological interventions, such as pet therapy, have emerged as a potential treatment for agitation symptoms. This literature review evaluates the application of  pet therapy and its efficacy on mitigating agitation in patients with Alzheimer’s. Synthesizing clinical trials, this paper focuses on the physiological and neurobiological mechanisms through which Animal Assisted Interactions (AAI) can reduce agitated behaviors. Mechanically, this paper explores how through observable decreases in physical stress indicators like pacing and verbal outbursts, as well as increases in measured biomarkers like oxytocin, AAI shows itself as a promising non-pharmacological alternative to treating agitation in patients. That being said, methodological limitations across current literature such as brief intervention timelines highlight the need for larger and highly controlled randomized controlled trials to establish long-term clinical protocols.

Introduction

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder that affects millions of individuals above the age of 40 globally and represents a major public health challenge. When most consider AD, cognitive decline is recognized as the hallmark symptom of the disorder. Despite this, it is often the non-cognitive manifestations—known as Behavioral and Psychological Symptoms of Dementia (BPSD)—that place the greatest burden on caregiving facilities and families of patients (Cerejeira, Lagarto, and Mukaetova-Ladinska 2012). Historically, agitation in AD patients has been managed through psychotropic medications; however, these pharmacological approaches carry significant risks, including sedative side effects and accelerated cognitive decline. Consequently, there is a clinical focus on identifying effective, non-pharmacological interventions (Cerejeira, Lagarto, and Mukaetova-Ladinska 2012).

 Animal-Assisted Interactions (AAI) have emerged as a compelling non-pharmacological therapy for use in geriatric care. AAI uses the human-animal bond to improve emotional, physiological, and social well-being (Filan and Llewellyn-Jones 2006). Therapy animals provide a non-verbal and non-judgmental companion whose contact with patients triggers a release of oxytocin along with a simultaneous decrease in cortisol levels and blood pressure (Kim et al. 2021).

Unfortunately, data on AAI for Alzheimer’s patients is severely hampered by methodological limitations, such as small sample sizes and insufficient follow-up tracking (Filan and Llewellyn-Jones 2006; Yakimicki et al. 2019).

Background

Alzheimer’s disease is a degenerative and irreversible neurological disorder that was first discovered in 1906 by German psychiatrist and neuropathologist Dr. Alois Alzheimer. It began when Dr. Alzheimer examined a middle-aged female patient, August Deter, who was exhibiting extreme and abnormal paranoia, memory loss, and agitation. Unable to prevent Deter’s death, Dr. Alzheimer performed a post-mortem histological examination hoping to find the cause for the observed deterioration of Deter’s physical, cognitive, and behavioral functions. Upon the examination, Dr. Alzheimer discovered the now hallmark signs of Alzheimer’s disease such as amyloid plaques and neurofibrillary tangles of brain fibers (Alzheimer 1907; Graeber and Mehraein 1999). These findings were presented to Dr. Alzheimer’s colleagues and the disease was coined “Alzheimer’s disease” nearly 4 years later in 1910 by Dr. Emil Kraepelin in his textbook Psychiatrie (Kraepelin 1910). At first, Alzheimer’s disease was believed to be a rare condition with an age barrier dictated by Deter who had developed the disease in her fifties (Graeber and Mehraein 1999; Kraepelin 1910). However, around the 1960s and 70s, what was originally thought to be “senile dementia” (cognitive decline that came with age) was found to be pathologically identical to Alzheimer’s disease (Blessed, Tomlinson, and Roth 1968). This discovery led to Alzheimer’s disease and dementia being interchangeable names for this cognitive condition that affects an overall population of around 55 million people worldwide (World Health Organization 2023) including an estimated 7.4 million Americans (Alzheimer’s Association 2024). The disease can show symptoms as early as in the patient’s 40s in rare cases of early-onset. However, most cases see a typical late onset of Alzheimer’s disease on individuals 65 or older (Alzheimer’s Association 2024). These diagnoses were gathered through measuring a patient’s cognitive decline as well as an objective framework based on living biomarkers (Jack et al. 2018). As part of that framework, amyloid-beta testing (brain imaging) and tau pathology testing (blood or cerebrospinal fluid tests) are used to establish the existence of abnormal amyloid plaques and the spread of toxic neurofibrillary tangles, respectively. The presence of these signs and symptoms are required to establish an Alzheimer’s diagnosis (Jack et al. 2018; Hansson et al. 2021).

Pharmacological intervention has been used for many years in an attempt to combat Alzheimer’s disease. Medicinal drugs, while impermanent, hold many benefits for patients (Cummings et al. 2022; Birks and Harvey 2018). Along with boosting neurotransmitters and managing behavior, some medicines can provide a temporary stabilization for Alzheimer’s patients that can help reduce caregiver burnout (Cummings et al. 2022; Birks and Harvey 2018). Even so, pharmacological limitations continued to halt scientists and doctors in developing a cure for Alzheimer’s (Cummings et al. 2022). Furthermore, pharmacological treatment can have adverse side effects such as gastrointestinal (GI) distress and increased risk of falling (Birks and Harvey 2018). Most of all, regular use of Alzheimer’s medication reduces its efficacy over time as more neurons die off (Cummings et al. 2022; Birks and Harvey 2018). As a result, modern medicine bifurcated its approach to the disease into two distinct focuses (Winblad et al. 2016). One track being long-term prevention and early detection of Alzheimer’s and the other being a focus on the management and comfort of patients already suffering from the disease (Winblad et al. 2016). With how progressive Alzheimer’s disease is, it became important to identify the warning signs of cognitive impairment before it developed into a near complete incapability of self-regulation (Jack et al. 2018, 535; Winblad et al. 2016). For those with Alzheimer’s disease, the primary clinical target of dementia care is quality of life (Winblad et al. 2016; Gitlin et al. 2012).

As such, there have been a wide variety of methods attempted ranging from pharmacological interventions to familial and personal caregivers (Gitlin et al. 2012, 2020). And while there are pharmaceutical treatments that manage to slightly slow cognitive decline, they often come with side effects and limited efficacy for behavioral symptoms (Cummings et al. 2022; Birks and Harvey 2018). This has prompted researchers to look for non-pharmacological therapies (Gitlin et al. 2012; Park et al. 2020). One such method that has recently gained more traction is pet therapy or animal assisted therapy (AAI) (Park et al. 2020). Pet therapy has already been proven to be effective in treating general mental health conditions like anxiety and depression (Charry-Sánchez et al. 2018). The presence of an animal companion helps many individuals with such afflictions to feel more grounded and calmer; and in hope of seeing a similar effect in patients with Alzheimer’s disease, pet therapy has gradually been introduced as a possible option for maintaining a reasonably high quality of life for individuals with Alzheimer’s (Park et al. 2020; Charry-Sánchez et al. 2018).

Effect of Pet Therapy on Agitation

The effectiveness of pet therapy in mitigating agitation is rooted in distinct psychological and physiological mechanisms. From a sensory perspective, the experience of stroking an animal’s fur returns the individual to the functions of the parasympathetic nervous system, increasing release of chemicals like acetylcholine and allowing calming hormones like oxytocin to take over. Furthermore, these positive interactions are associated with reduced production of the stress hormone cortisol, which was detected in saliva samples from patients (Sheikh et al., 2021). Additionally, what Sheikh’s literature review found that is especially promising about AAI is the bi-directional aspect that can be seen with pets such as dogs that dynamically react to human emotion — for example, when a dog holds eye contact with a human, or when a dog knows to lean into a person to help them feel accepted by the animal.

Managing the behavioral and psychological symptoms of dementia (BPSD)—particularly agitation, aggression, and late-afternoon distress known as sundowning—presents a profound clinical challenge in Alzheimer’s care for both the caregivers and those with Alzheimer’s. Clinical trials within high-stress medical settings demonstrate that the introduction of a trained therapy dog yields statistically significant, immediate reductions in patient agitation compared to standard care routines (Krüger et al., 2021). Kruger’s study was a systematic review of 51 studies concerning AAI with patients with Alzheimer’s disease. The study found that there was a correlation between animal intervention and a decrease in agitation, while oppositely finding no significant relationship between AAI and depression/quality of life.  Even so, this behavioral shift suggests that the physical presence of an animal provides a grounding, non-threatening stimulus capable of redirecting a patient’s attention away from environmental stressors or internal confusion.

Despite these positive outcomes, a comprehensive analysis of the literature reveals critical methodological limitations regarding the sustainability of these interventions. While the immediate de-escalation of agitation during a pet therapy session is well-documented, evidence tracking the long-term or residual effects after the animal leaves remains scarce and often low-certainty (Park et al. 2020). Park performed a systematic review of 9 randomized controlled trials published between 2000 and 2019. It was found that there is a statistically significant amount of evidence that AAI reduces depression in dementia patients. However, with broader outcomes like sustained agitation reduction and long-term behavioral changes, the overall evidence base was limited by small sample sizes and a lack of follow-up tracking. Additionally, individual patient histories must be considered; for individuals with pre-existing cynophobia (fear of dogs) or animal allergies, the introduction of a therapy animal can inadvertently exacerbate psychological distress rather than soothe it. Consequently, current research positions pet therapy as a highly potent, acute tool for situational agitation management, though further standardized, longitudinal studies are required to determine how to extend these calming effects into a patient’s broader daily routine.

Research Gaps 

Most studies concerning AAI in use for treating stress and agitation in dementia patients all share similar research limitations. Those limitations being that oftentimes the studies lack sufficient sample sizes or long-term follow-up for the realistic viability of using therapy animals in caregiving settings — to address this, prospective longitudinal studies should be conducted. Another variable these studies fail to capture is the maintenance aspect of using therapy animals to help treat dementia patients. The care for the therapy animals adds another layer of work to be done for patients that already require around the clock care. Even if there were to be specific people assigned to taking care of the therapy animals, it would require more people in the caregiving/nursing profession which is already short staffed.

Conclusion

The synthesis of current literature indicates that pet therapy can serve as a valuable non-pharmacological strategy for reducing agitation and behavioral distress in individuals living with Alzheimer’s disease. The presence of therapy animals consistently yields reductions in acute agitation, accompanied by increased positive emotional engagement. The physical presence provided by animals appears to mitigate anxiety without the adverse risks often associated with psychotropic medications.

However, while short-term behavioral benefits during active sessions are well-supported, long-term durability remains questionable. Future research should prioritize larger sample sizes and designs that allow researchers to follow up with patients from the sample long after the end of the initial experiment. In this way, the research on AAI with Alzheimer’s patients can hopefully be expanded and possibly conformed into a standardized protocol in caregiving facilities.

References

Alzheimer, Alois. 1907. “Über eine eigenartige Erkrankung der Hirnrinde.” Allgemeine Zeitschrift für Psychiatrie und Psychisch-Gerichtliche Medizin 64: 146–148.

Alzheimer’s Association. 2024. “2024 Alzheimer’s disease facts and figures.” Alzheimer’s & Dementia 20 (5): 3708–3821.

Birks, Jacqueline S., and Richard J. Harvey. 2018. “Donepezil for dementia due to Alzheimer’s disease.” Cochrane Database of Systematic Reviews, no. 6.

Blessed, G., B. E. Tomlinson, and M. Roth. 1968. “The association between quantitative measures of dementia and senile change in the cerebral grey matter of elderly subjects.” The British Journal of Psychiatry 114 (512): 797–811.

Cerejeira, Joaquim, Luis Lagarto, and Elka B. Mukaetova-Ladinska. 2012. “Behavioral and psychological symptoms of dementia.” Frontiers in Neurology 3: 73. https://doi.org/10.3389/fneur.2012.00073.

Charry-Sánchez, Juan D., et al. 2018. “Animal-assisted therapy in adults: A systematic review.” Complementary Therapies in Clinical Practice 32: 169–180.

Cummings, Jeffrey, et al. 2022. “Alzheimer’s disease drug development pipeline: 2022.” Alzheimer’s & Dementia: Translational Research & Clinical Interventions 8 (1): e12295.

Filan, S. L., and R. H. Llewellyn-Jones. 2006. “Animal-assisted therapy for dementia: A review of the literature.” International Psychogeriatrics 18 (4): 597–611. https://doi.org/10.1017/s1041610206003322.

Gitlin, Laura N., et al. 2012. “Nonpharmacologic management of behavioral symptoms in dementia.” JAMA 308 (19): 2020–2029.

Graeber, Manuel B., and P. Mehraein. 1999. “Reanalysis of the first case of Alzheimer’s disease.” European Archives of Psychiatry and Clinical Neuroscience 249 (3): S10–S13.

Hansson, Oskar, et al. 2021. “The Alzheimer’s Association appropriate use recommendations for blood biomarkers in clinical practice.” Alzheimer’s & Dementia 17 (10): 1675–1693.

Jack, Clifford R., et al. 2018. “NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease.” Alzheimer’s & Dementia 14 (4): 535–562.

Kim, S., Y. Nam, M.-J. Ham, C. Park, M. Moon, and D.-H. Yoo. 2021. “Neurological mechanisms of animal-assisted intervention in Alzheimer’s disease: A hypothetical review.” Frontiers in Aging Neuroscience 13: 682308. https://doi.org/10.3389/fnagi.2021.682308.

Kraepelin, Emil. 1910. Psychiatrie: Ein Lehrbuch für Studierende und Ärzte. 8th ed. Johann Ambrosius Barth.

Kruger, Katelyn, et al. 2021. “Animal-Assisted Intervention in Dementia.” ResearchGate. https://www.researchgate.net/publication/315461366_Animal-Assisted_Intervention_in_Dementia. 

Park, A. L., et al. 2020. “Animal-assisted interventions for people with dementia: A systematic review and meta-analysis.” Journal of Clinical Nursing 29 (17-18): 3121–3133.

Sbrizzi, C., and W. Sapuppo. 2021. “Effects of pet therapy in elderly patients with neurocognitive disorders: A brief review.” Dementia and Geriatric Cognitive Disorders Extra 11 (3): 198–206. https://doi.org/10.1159/000518469.

Sheikh, Saba, et al. 2021. “Pet-Assisted Therapy for Delirium and Agitation in Hospitalized Patients with Neurocognitive Impairment: A Review of Literature.” Geriatrics 6 (4): 96. doi.org.

Winblad, Bengt, et al. 2016. “Defeating Alzheimer’s disease and other dementias: a priority for European science and society.” The Lancet Neurology 15 (5): 455–532.

World Health Organization. 2023. Dementia key facts. World Health Organization.

Yakimicki, Mary L., Nancy E. Edwards, Elizabeth Richards, and Alan Beck. 2019. “Animal-assisted intervention and dementia: A systematic review.” Clinical Nursing Research 28 (1): 9–29. https://doi.org/10.1177/1054773818756987.

About the author

Colin Weindel

Colin is a senior at Carmel High School who is hoping to pursue a career in veterinary medicine after graduating high school starting with attending university to major in Animal Science.

The post A Literature Review of the Effect of Animal-Assisted Therapy in Reducing Agitation in Patients with Alzheimer’s Disease appeared first on Exploratio Journal.

Introduction and argument

In Thucydides’ History of the Peloponnesian War, the Athenians argue that “three of the strongest motives” in international affairs are “fear, honor, and interest” (Thucydides, I.76). The Cold War followed Thucydides’ pattern, though it never erupted into direct US-Soviet military conflict. Each superpower feared its adversary’s global power, desired honor, and recognition, and pursued its geopolitical interests. Together, these three motives explain both the persistence and the “coldness” of the conflict, showing that Thucydides’ ancient framework remains relevant to the recent nuclear armed conflict. 

Section 1: Fear of Adversaries Global Power 

A: Communism Concern by the United States – A connection to ideologies 

American fear during the Cold War was rarely a fear of Soviet troops alone; it was rather a fear of communism as an ideology spread worldwide without direct invasion. This pattern echoes the dynamics before the outbreak of the Peloponnesian War.  The Athenian and Spartan influence going outward through allied city states is more than a simple fear of conquest, like the Cold War. (Gaddis, 1997). 

George Kennan first articulated his theory in his February 1946, “Long Telegram” and his July 1947 “The Sources of Soviet Conduct,” published anonymously under “X” in Foreign Affairs. Kennan attributed Soviet behavior to Marxist-Leninist ideology and a long-standing Russian strategic culture, arguing that the Kremlin was ideologically committed to the belief that capitalism was doomed to collapse before the Soviet state would (Kennan, 1947). Kennan called for a “long-term, patient but firm and vigilant containment of Russian expansive tendencies” (p. 581) through a policy of political and economic pressure, not military confrontation. 

In April 1950, Paul Nitze drafted NSC-68, a policy paper that succeeded Kennan at the State Department’s Policy Planning Staff. Kennan proposed confronting Soviet power at strategic points. (Gaddis, 2005). NSC- 68 reimagined the Soviet threat as a near-total civilization danger, describing the conflict as one between a “free society” and system “so wholly irreconcilable with ours, in its purpose to destroy ours “(National Security Council, 1950). Gaddis argues that this shift reflected a growing conviction that the US could not allow challenges to the existing global order even to appear to succeed (Gaddis, 2005). This maps onto Thucydides’ theory of fear. For both Sparta and the Cold War superpowers, it was driven less by an adversary’s current capabilities than the trajectory of its rise. In the Cold War, the consequence was a heavy increase in military spending, from $13 billion (about $40 per person in the US)13 billion (about $40 per person in the US) in the 1950 budget to over 60 billion by 1951. 

B:  Technological Development: A race to the top 

Arguably, no single event solidified American fear of Soviet capability more than the launch of Sputnik on October 4^th, 1957. The launch from the Baikonur Cosmodrome produced what Dwight D. Eisenhower later called a “wave of near hysteria” in the American press and public eye. (Moody, 2020) The New York Times ran 279 articles mentioning Sputnik in three weeks following the launch (Moody, 2020). The striking part about this episode is how disproportionate the reaction was to Sputnik itself – it wasn’t the physical object itself that imposed fear on the public, it was merely a suggestion that if the Soviets could get this into object into space, then surely, they could also land a nuclear warhead on American soil. This fear was instilled upon and nurtured into the American psyche, people, and government. Senator Lyndon Johnson’s six weeks of Senate hearings on the matter eventually led to new federal institutions: US Congress passed the National Aeronautics and Space Act in July 1958, creating NASA alongside the National Defense Education Act, which restructured American math and science education around the technological contest. NASA was created to give the American program a civilian and scientific face to compete with the Soviet Union’s technological power. (McDougall, 1985) 

C: Fear of Nuclear War 

The mutual fear of nuclear war is the strongest evidence that Thucydides’ account of the Peloponnesian War applies to US-Soviet relations during the early Cold War. For Thucydides, Spartan fear of Athens’ growing power was the leading cause of the war itself. In the Cold War, the same dynamic existed between Washington and Moscow, but the two powers did not fight each other directly. By the late 1950s and increasingly through the 60s, both sides had the capacity to assure destruction of the other, making direct war between the superpowers irrational for either side. This did not quell fear; if anything, it intensified with each new weapon system.  Instead of escalating towards direct physical confrontation, fear was placed where the risks fell short of triggering any nuclear conflict. 

Fear made direct war over the Central European front unthinkable, both global superpowers redirected their energy towards asserting influence in a decolonizing world of new uncommitted governmental states. Influence could be won or lost but needed not trigger the catastrophe both sides most feared: nuclear war. 

Section 2: Honor, Desire for Recognition by both sides 

Honor is the second of Thucydides’ three motives, and while the Soviet drive for honor is the most discussed phenomenon, it was not purely one sided. The concept of American honor deserves equal attention and with the case of The Marshall Plan it illustrated that the United States was pursuing prestige almost as intensely as Stalin. These two simply look different because one side had already been recognized as a leading power, while the other was trying to become one. 

A: American Honor and the Marshall Plan 

Secretary of State George Marshall announced the European Recovery Program at Harvard University on June 5, 1947. There, he insisted that “our policy is directed not against any country or doctrine but against hunger, poverty, desperation, and chaos,” (Marshall, 1947). The framing was intentional and should not be dismissed as rhetoric, because it was through this intentional lens that American officials attempted justify their plan suggesting that other things were at stake beyond economics. Historian Michael Hogan’s definitive study of the program argues that the Marshall Plan was never simply an effort to put American aid behind European reconstruction. It was an attempt to project a distinctly American model of integrated capitalism and democracy onto Western Europe and to demonstrate that American leadership could produce results that Soviet-style central planning could not. The plan was just as much about what the United States wanted the world to think of it. If the Marshall Plan was purely ‘interest-based’ then the Marshall Plan struggles to explain why the United States insisted that Europe jointly designed and administered the programs themselves rather than simply just dictating the terms the way a state acting in self-interest alone might have. In its existence on partnership and not just aid, it looks like the claim about a character of American power, not merely its size, highlighting the point on honor. 

Benn Steil’s recent account, “The Marshall Plan: Dawn of the Cold War,” pushes this argument even further, arguing that the plan should be understood as the true beginning of the Cold War precisely because it was the moment where both superpowers became committed to demonstrating the superiority of their respective systems to the rest of Europe not merely just to securing their own security. (Steil, 2018) This framing is persuasive because it explains a detail that is purely strategic account leaves unexplained: the Marshall Plan was formally offered to the Soviet Union and Eastern European satellites as well. Molotov’s walk out from Paris, planning talks in July 1947, followed by Stalin’s refusal to allow any Eastern European state to participate, was a choice to protect Soviet prestige. (Steil, 2018) his offer and refusal, in other words, were both honor-inflected moves. The United States wanted credit for extending the offer as much as it wanted the offer accepted and the Soviet Union wanted to be seen as rejecting American charity on its own terms rather than being seen to need it. 

Steil argues that within Western Europe the plan succeeded in establishing durable American institutional leadership, which laid groundwork that will later show NATO in the broader post-war order precisely because it was designed to be generous and cooperative rather than just merely transactional. (Steil, 2018) 

This is, if not the clearest evidence that the United States was pursuing honor alongside interest in this period. A purely self-interested aid program did not need to be this publicly framed as moral leadership, in the fact that it suggests American policymakers cared how the gesture would be read, not just what they buy. 

B: UN Recognition, Stalin’s demand for veto power 

If the Marshall Plan showed American honor, the founding of the United Nations shows Soviet honor expressed through the opposite impulse: an insistence on formal procedural equality that no amount of goodwill could substitute for the United Nations. This offers the earliest and clearest Cold War illustration of honor operating as a distinct motive from fear or interest. 

At the 1944 Dumbarton Oaks Conference and the February 1945 Yalta Conference, the Soviet delegation insisted that each of the great powers possess an absolute veto over the Security Council action. In other words, as a permanent member, this veto power was broad enough to block even the discussion of a dispute, a position the British delegation resisted on the grounds that no state should be able to veto a matter to which it was a party (Luck, 1995, p. 509-539). Stalin personally justified this position to Roosevelt by offering that unanimity among the four leading powers, which included the USSR, the United States, the United Kingdom, and China, was necessary to prevent the great powers from being placed in opposition to one another. Now most historians would argue, including Geoffrey Roberts, that the underlying motive ran deeper than procedural caution (Roberts, 2008). The veto was, in Roberts’ framing and in a broader post-war Soviet self-conception, a symbol of the USSR’s arrival as a global power after two decades of being treated by the Western powers as an ideological pariah state following the 1917 Bolshevik revolution (Roberts, 1999, p. 655-673). 

Territorial gains at the war’s end mattered to Stalin not only for physical security but also because they underwrote the USSR’s claim to sit as an equal among victors in shaping post- war order (Judt, 2006). 

How the Soviet Union used its veto power it possessed makes the honor motive explicit. In the United Nations’ first decade, the USSR was the only permanent member to invoke the veto at all, doing so 57 times by 1955 (Luck, 1995). This rate of unilateral obstruction reflected judgment on each resolution, but a generalized defensiveness against the Security Council that the Kremlin perceived. 

By fall of 1947, the Soviets had cast 17 of 18 vetoes issued in the council’s brief history (Gaiduk, 2006). What that pattern suggests is that the Soviet UN policy during this period is better read as honor-protective than purely interest-maximizing. This veto was used less to secure one specific outcome but to prevent the council from rendering judgment that would cast the USSR as a lesser power when Moscow felt disrespected. 

Its refusal to accept the nationalist Chinese delegation’s continued occupation of China’s seat which led to its refusal to accept the United Nations Security Council altogether from January to August of 1950. This decision inadvertently allowed the council to authorize UN intervention in Korea in the USSR’s absence (Mastny, 1996). The outcome is clear as a demonstration that honor and interest can pull in opposite directions. A state acting purely on interest would never have vacated its seat when its veto was most needed, but a wounded state did precisely that. 

C. The Soviet Atomic Bomb as a Marker of Great Power Status 

Just as Sputnik sparked a U.S Space Race, no episode of the early Cold War illustrates the honor motive more directly than the Soviet drive to build an atomic bomb. David Holloway’s prize-winning “Stalin and the Bomb” (still a definitive academic treatment of the subject) shows that Stalin authorized a nuclear program immediately after the American bombings of Hiroshima and Nagasaki in August of 1945. The American atomic monopoly shaped nearly every aspect of Soviet foreign policy in their intervening four years (Holloway 1994). Holloway’s account is explicit that the bomb was never purely a military project for Stalin, but it was a system where priority was given to a high-technology program only once it was known the capitalist West was pursuing the same technology. The atomic project sat alongside rocketry, jet engines, and radar as a test of whether the Soviet Union could match Western capability on the most prestigious frontier (Holloway 1994). The American atomic monopoly between 1945 and 49, in this sense, was a standing insult to the Soviet claim to be a genuine equal among the victorious wartime powers. 

Holloway describes Stalin and Soviet officials as “visibly elated” when the USSR successfully tested their first device, code named “the First Lightning,” modeled after the American “fat man” design at the Semipalatinsk site in Kazakhstan on August 29, 1949. This reaction makes little sense if the bomb was purely ‘strategic’ since the Soviet Union still lacked any reliable means of delivering a weapon against the American mainland (Holloway, 1994). What the test did deliver was status and within days of the test, President Truman was forced to publicly confirm that an atomic explosion had occurred in Soviet territory, which ended the four years where nuclear weapons had been the exclusive marker separating the two emerging superpowers from every other state on Earth (Holloway, 1994). 

Historian Martin Sherwin’s influential American Historical Review article on the origins of American atomic diplomacy says that the US officials had themselves treated the bomb as a diplomatic instrument from the outset, signaling it’s use to Moscow even before the war against Japan ended. This makes the Soviet achievement of 1949 a direct and calculated answer to an instrument of American prestige, not merely a response to a military threat. (Sherwin, 1973. p. 566-582). This reads onto Thucydides: the episode shows honor operating exactly as it does in the Thucydidean Peloponnesian War. 

Section 3: Interest  

A: The East and West Germany Conflict 

Section 1 centers around fear pushing these two superpowers away from direct confrontation, Section 2 is where that displaced energy went – no case study makes that clearer than Germany. Germany arguably was the single and clearest site of honor driven competition than in the entirety of the early Cold War because it was the one place where territory, ideology and hard security all converged into a single space rather than being spread across separate crises. The 1948-49 Berlin Blockade is the founding case study. After three Western occupying powers announced a currency reform introducing the Deutscher Mark into their zones, Stalin ordered a blockade of all rails, road and canal access to the Western sections of Berlin, an island of Allied control deep inside the Soviet occupation zone (Shlaim, 1983). Stalin’s aim was not primarily punitive; it was interest-maximizing in the most literal Thucydidean sense, an attempt to force a Western withdrawal and consolidate exclusive Soviet control over the entirety of the city and, by extension, leverage over the shape of a future German state (Harrington, 2012) The Western response, the Berlin Airlift, sustained the population of West Berlin by air for nearly a year. Historian Harrington’s archival study demonstrates the crisis originated much earlier than the blockade itself. In wartime occupation planning that never resolved basic questions of access and sovereignty, running through the failed 1949 Council of Foreign Ministers meeting in Paris (Harrington, 2012). That earlier origin point matters to the argument because it means the blockade was not a sudden Soviet decision so much as the eruption of ambiguity that had existed since 1945. Interest-driven competition does not need a single event when underlying stakes were unresolved from the start. Harrington’s account is also useful because it resists a purely top-down narrative that credits the operational ingenuity of pilots, mechanics, and ordinary Berliners as much as high-level diplomacy for the crisis’s peaceful solution. (Harrington,2012). 

The Berlin question resurfaced with higher stakes in 1958, when Khrushchev issued an ultimatum demanding Western withdrawal from Berlin within six months, and this culminated the construction of the Berlin Wall in August of 1961. (Stivers, 1997, p. 569- 602) As Thomas Schwartz’s scholarship on the crisis suggests, the episode produced a durable, if unspoken, mutual understanding. The Soviets wanted to continue to dominate Eastern Europe and Eastern Berlin, while the United States and allies would retain their sphere over Western Europe, West Germany, and West Berlin (Schwartz, 1997. p. 945-968) The outcome is best read through the lens of Thucydides; Melian Dialogue, in which Athenian envoys tell the Melians that “the strong they do what they can and the weak suffer what they must.” That comparison must be careful with understanding, because the Cold War Germany was not a case of coercion the same as Melos, but the underlying logic is the same: neither superpower could tolerate the slightest appearance of losing ground in Germany without risking a cascading loss of credibility and honor throughout the rest of Europe. That is the real argument for why interest and reputation cannot be separated in this case. Every local outcome in Berlin was read by both sides as a signal of the broader rivalry, which is exactly the kind of relative, calculation Thucydides describes in both Athens and Sparta. 

B: Early Soviet Control over Eastern Europe 

If Berlin was the most visible flashpoint of interest-driven competition, then the Soviet consolidation of Eastern Europe between 1945 and 1948 was the broadest and, in sense, its most honest expression. Anne Applebaum, prize winning historian argues in The Iron Curtain that Sovietization of the region was not primarily a defense action to Western provocation, as some other historians claim, but a methodical and deliberately planned extension of Soviet interest that began well before the Cold War officially started (Applebaum, 2012). The timing of this argument is important because it pushes back on any of the reading of the early Cold War as a story of mutual, symmetrical escalation; on Applebaum’s account, one side had already decided what it wanted before the other side fully reacted. Communist parties that were originally installed in broad coalition governments across Poland, Hungary, Romania and Bulgaria steadily eliminated their rivals through intimidation, rigged elections, and direct interventions by Soviet forces, so that by 1948 every state from the Baltic to the Balkans with the exception of Greece and occupied Austria had been absorbed into the Soviet Sphere (Applebaum, Iron Curtain). This is interest in its most territorial and least disguised form; Stalin was not competing for influence at a distance, as the superpowers later would in other settings, but converting military occupation into permanent political control. 

Vojtech Mastny’s The Cold War and Soviet Insecurity offers a complimentary explanation for why Stalin pursued his policy so quickly, an abiding and almost obsessive Soviet anxiety about the security following the catastrophic losses of the Second World War, in which roughly 27 million Soviet citizens died. (Mastny, 1996). This is worth taking seriously rather than dismissing as a justification, because it changes how one should categorize the episode:  Eastern Europe functioned as a buffer zone; intended to prevent any future invasion into Russia from the west, giving Soviet interest in the region as defensive, as well as expansionist character; though, that distinction mattered little to the people of Warsaw, Budapest of Prague living under one-party regimes. Reading Melvyn Leffler’s, A Preponderance of Power, alongside Mastny emerges less of a story of one side pursuing interestand the other reacting than two genuinely separate interests; Soviet Security and American credibility, which happened to be mutually exclusive in the same geographic space. Read through a Thucydidean lens, the Sovietization, tying it back to Applebaum’s argument, illustrates interest operating almost exactly as it does in Thucydides’ argument. The Soviet Union had the capacity to regulate outcomes in the region that it ruled or occupied, while the interests of the populations themselves were unconsidered in Moscow’s calculations (Leffler, 1992) and because of the way these governments were installed into other countries, the Yalta conference promise of free elections became the central American grievance in the recognition disputes of 1945 and 1946. (Applebaum, 2012) 

The process carried a dimension of honor along with a material one. At the February 1945 Yalta Conference, the US, Britain, and the Soviet Union signed the Declaration on Liberated Europe, promising to help liberated people to form governments “through free and unfettered elections” (Leffler, 1992). Washington related Soviet non-compliance not only as a security setback but as a breach of Stalin’s promise among equality. Stalin, for his part, regarded Western protests as fundamentally hypocritical, given that the US exercised comparable control over post was settlements in Japan and Italy without inviting Soviet objection (Mastny, 1996). Each side of being disrespected fed into others showing that even being territorial, Soviet interested in Eastern Europe could not be fully separated from status and legitimacy. 

Section 4: The Limits of the Analogy 

The evidence presented in Sections 1-3 shows that fear, honor and interest all shaped US – Soviet relations in ways that closely track Thucydides’ ideology, however, the fit is not always perfect and acknowledging where it strains is necessary to answer the “to what extent” portion of the question rather than accumulating favorable examples.  

A: Fear inversion 

The clearest break from Thucydides lies in what fear produces during these diplomatic relations. For Thucydides, Spartan fear of Athens’ rising power was the efficient cause of war; fear escalated directly into conflict for him. In the Cold War, the opposite held: once both powers possessed the capacity for mutually assured destruction, fear made direct war irrational rather than likely. This is not a minor variation on the same mechanism, and it is arguably an inversion of it. The triad explains why states compete, but nuclear weapons changed what competition could escalate into. The triad was built on a world without weapons capable of mutual annihilation and cannot, without modification, explain why the two most powerful and most afraid states in modern history never directly fought each other. The triad survives the Cold War only if fear is redefined to mean that something closer to risk- adverse status of competition than the war producing fear. 

B: Ideology has no Thucydidean equivalent  

Section 1A noted that the American fear of Soviet Union was a fear of communism as an overarching idea, not just territorial power as Athenian influence was directed through allied city states rather than a simple fear of conquest. This comparison understates the difference; Athenian influence over its allies was maintained through tribute and naval power- much more materialistic tools. American fear of communism and Soviet fear of liberalism involved a belief that the system itself was contagious and would spread through voluntary conversion, not conquest (Kennan, 1947). In this case, Thucydides’ idea had no equivalent concept, because no state in his time claimed a universal ideological mission the way that the Cold War powers did. This suggests that the triad explains the Cold War’s competitive structure quite well but requires the additional variable of ideological universalism. 

C: Two Superpowers, not as many City States 

A smaller but real limitation: Thucydides describes a world of city states and shifting alliances, where fear, honor and interest operate relationally among many actors. The Cold War’s bipolarity skews this calculation. When two states possess overwhelming power, the contest for honor like the UN veto or Marshall plan is not one competition move, but a signal read by the entire world as evidence for one side or the other. This concentrates the stakes of every individual episode in a way that Thucydides’ idea did not require (Thucydides, 1998). 

Why does the framework still hold, with modification

None of these points disprove the paper’s central argument. Fear, honor, and interest remain in the three organizing moves of US- Soviet relations; the evidence across Sections 1 through 3 makes that argument convincing. What nuclear weapons and ideology changed was not whether these three motives operated, but how they were expressed at the time and what they eventually would escalate into. Thucydides’ triad describes the grammar of great power escalation; the Cold War supplies the vocabulary to the grammar, by conditions no ancient historian could have remotely anticipated. 

Conclusion 

All three motives were present throughout the early Cold War, but their expressions shifted under the shadow of nuclear weapons. The fit is not seamless; nuclear deterrence inverted the mechanism that Thucydidean fear produced in war, and neither ideological universalism nor rivalry had any precedent in the ancient world. Fear, rather than escalating toward war as in Thucydides’ Greece, discharged itself into competition over interest, such as Germany, Eastern Europe, etc., and honor, the UN veto, the atomic bomb, recognition disputes. Both methods intertwined with interest while also playing a role of their own. This suggests that the early Cold War does not disprove Thucydides’ triad. It only reveals its adaptability to conditions that he could not predict. 

Works Cited 

Applebaum, A. (2012). Iron curtain: the crushing of Eastern Europe, 1944-1956. Doubleday. 

Gaddis, J. L. (1997). We now know rethinking Cold War history. Oxford University Press. 

Gaddis, J. L. (2005). The Cold War: a new history. Penguin Books. 

Harrington, D. F. (2012). Berlin on the brink: the blockade, the airlift, and the early Cold War. University Press of Kentucky. 

Hogan, M. J. (1987). The Marshall Plan: America, Britain, and the reconstruction of Western Europe, 1947-1952. Cambridge University. 

Holloway, D. (1994). Stalin and the bomb: the Soviet Union and atomic energy, 1939-1956. Yale University Press. 

Judt, T. (2006). Postwar: a history of Europe since 1945. Penguin Books. 

Kennan, G. F (1946, February 22). The long telegram. In Foreign relations of the United States, 1946 (Vol. VI) U.S Department of State, Office of the Historian https://history.state.gov/historicaldocuments/frus1946v06/d475 

Kennan, G. F. (writing as “X”). (1947, July). The sources of Soviet Conduct. Foreign Affairs. https://www.foreignaffairs.com/russian-federation/george-kennan-sources-soviet-conduct 

Leffler, M. P. (1992). A preponderance of power: national security, the Truman administration, and the Cold War. Stanford University Press. 

Luck, E.C. (1995) The Security Council’s First Fifty Years, American Journal of International Law 89(3), 506–539. 

Marshall, G. C. (1947, June 5). The Marshall Plan speech [Speech transcript]. George C. Marshall Foundation  https://www.marshallfoundation.org/life-legacy/

Mastny, V. (1996). The Cold War and Soviet insecurity: the Stalin years. Oxford University Press. 

McDougall, W. A. (1985). The Heavens and the Earth: A Political History of the Space Age. Basic Books. 

Moody, S.J (2020). The view from NATO: Sputnik as a catalyst, 1957-8. The International History Review, 42(6)  

National Security Council. (1950, April 7) NSC-68: United States objectives and programs for national security. U.S Department of State, Office of the Historian https://history.state.gov/milestones/1945-1952/NSC68 

Roberts, G. (1999). Ideology, calculation, and improvisation: Spheres of influence and Soviet foreign policy, 1939-1945. Review of International Studies, 25(4), 655-673 

Roberts, G. (2008). Stalin’s wars: from World War to Cold War, 1939-1953 (Pbk ed.). Yale University Press. 

Schwartz, T.A. (1997). The Berlin crisis and the Cold War. Diplomatic History, 21(1), 139-148. 

Sherwin, M. J. (1973). The Atomic Bomb and the Origins of the Cold War: U.S. atomic-energy policy and diplomacy, 1941-45. American Historical Review, 78(4), 945-968. 

Shlaim, A. (1983). The United States and the Berlin Blockade 1948-1949: a study in crisis decision-making. University of California Press. 

Steil, B. (2018). The Marshall Plan: dawn of the Cold War. Simon & Schuster Paperbacks. 

Stivers, W. (1997). The incomplete blockade: Soviet zone supply of West Berlin, 1948-1949. Diplomatic History, 21(4), 569-602 

Thucydides. (1998). History of the Peloponnesian War (R. Crawley, Trans.). Wordsworth Editions. (Original work published ca. 431 BCE)  

About the author

Ryann Weinstein

Ryann is a senior at Jupiter Community High School in Jupiter, Florida. In her free time, Ryann loves to read, especially History, and travel. Ryann has been on numerous service adventures where she has traveled to Bali, Greece and Costa Rica to work with children, help environmentally, and immerse herself in culture.

Her love of travel has inspired Ryann to pursue international relations in college and eventually fulfill the dream of working in the foreign service.

The post To what extent does Thucydides’ theory of fear, honor and interest present in US – Soviet relations during the early Cold War? appeared first on Exploratio Journal.

Abstract

India’s Unified Payments Interface (UPI) has reshaped the country’s financial landscape, establishing India as the global leader in real-time digital payment adoption. However, this rapid expansion has been accompanied by a corresponding rise in digital payment fraud, raising concerns for consumer protection and the effectiveness of structural systems. Applying Routine Activity Theory, the paper analyzes literature to explain this outcome as the product of a convergence between motivated offenders, suitable targets, and absence of capable guardians. Recent studies show that victimization remains consistent despite awareness of digital payment frauds. Educated and technologically aware users continue to fall victim to complex psychological manipulation. Once targeted, victimization largely depends on the exploitation of cognitive biases. Institutional guardianship remains India’s most persistent weak point. Victims recover a fraction of the stolen funds, redressal mechanisms are largely unsatisfactory, and legal loopholes leave a majority of victims outside of the protection system. The paper concludes with recommendations to put a stop to legal loopholes, modernize legal frameworks, and add an additional layer of protection at the point of transaction rather than relying solely on user vigilance.

Introduction

India’s creation and adoption of the Unified Payment Interface (UPI) is an example of one of the most rapid financial digitisation initiatives in recent history, expanding from 20 million transactions in  FY 2016-17 to over 241.62 billion in FY 2025-26 (Ministry of Finance, 2026). This represents a 12,000-fold increase.  In 2012, over 87% of all transactions in India were conducted in cash, and fewer than 10% of the population had ever used any non-cash payment instrument (Iyer, 2017). Today, India accounts for 48.5% of worldwide real-time digital payment transactions, making it the global leader in digital payments (Singh and Agarwal, 2025). Furthermore, of the total retail digital payment transactions, UPI accounted for over 81% of them, emerging as the largest real-time retail payment system globally (Ministry of Finance, 2025). Between April 2016 (at launch) and March 2026, the number of banks live on UPI increased from 21 to 703. These statistics prove that UPI has evolved from merely being a means of payment to a financial infrastructure critical to India’s economy.

However, the same expansion that produced this success also generated a fraud crisis of proportionate scale. According to the Association of Certified Fraud Examiners, fraud is any activity that relies on deception in order to achieve a gain. Fraud becomes a crime when it is a knowing misrepresentation of the truth or concealment of a material fact to induce another to act to his or her detriment. Within the UPI ecosystem, fraud commonly involves social engineering and other deceptive practices that manipulate users into authorising transactions or disclosing sensitive credentials. The number of reported UPI fraud cases rose from 195,000 in 2020-21 to 7,25,000 in 2022-23, while the amount lost increased from ₹111 crore (approximately $15 million as per the exchange rates for FY 2020-21) to ₹573 crore (approximately $72.4 million as per the exchange rates for FY 2022-23) in the same period (Agrawal, 2026). Across all digital payment modes, fraud cases increased from 1,19,699 in 2020-21 to 14,57,000 in 2023-24, recording a compound annual growth rate (CAGR) of 130.03%, which outpaces the 40-105% annual growth in UPI transactions (Kumar and Dharshan, 2026). Iyer (2017) documented that card fraud in India increased in direct proportion to card adoption throughout 2011–2016, with losses doubling from ₹4 billion ($89.58 million as per the 2011 exchange rates) to ₹7.68 billion ($113.56 million as per the 2016 exchange rates) as card usage expanded. Moreover,  according to Wadkar et al. (2025), UPI fraud accounts for 47.25% of all cybercrime cases reported in India since 2020.Understanding why fraud has scaled at this rate requires examining not only the architecture of UPI itself, but the behavioral and structural conditions that enable exploitation. Statistics alone cannot explain why UPI fraud has scaled at this rate or why users remain vulnerable despite high awareness. This paper applies Routine Activity Theory (Cohen and Felson, 1979) to examine the structural and behavioral conditions that enable this crisis.

Theoretical Framework:

According to Routine Activity Theory (Cohen and Felson, 1979), a criminal event occurs when three elements converge in both time and space: a motivated offender who perceives an opportunity for gain, a suitable target, and the absence of a capable guardian who can prevent the crime. Additionally, Cohen and Felson argue that changes in routine activities, such as shifts in work patterns, leisure activities, and household arrangements, impact the convergence of motivated offenders, suitable targets, and the lack of capable guardians. This, in turn, affects crime rates. Scholars have since extended RAT beyond street crime to fraud contexts. Holtfreter et al. (2008) conducted a study that extended the Routine Activity Theory (RAT) to the context of consumer fraud victimization,  using a sample of 922 participants. Their study found that a one-unit increase in the buying routine index (a measure of how frequently respondents engaged in consumer activities such as online shopping) was associated with a 61% increase in the odds of being targeted for fraud.

Holtfreter et al. (2008) also tested Self-Control Theory alongside RAT, finding that while low self-control significantly increased the probability that a targeted individual would become a victim (302% per unit increase), it had no effect on whether the individual was targeted in the first place. This targeting/victimization distinction is directly relevant to the UPI fraud paradox examined later in this paper: why do 52.5% of users become victims despite 96.2% awareness? Once targeted, a user’s capacity for deliberate, cautious decision-making (rather than prior knowledge alone) determines victimization outcomes.

DeLiema (2018) reinforces this by applying RAT to study financial fraud against older adults, examining 53 cases of financial exploitation. The study investigated interpersonal financial exploitation amongst adults aged 65 years or older. The research found that social isolation, which means a lack of capable guardians like trusted friends, family, or oversight from institutions, was the main factor allowing these crimes to happen. The study also highlighted that perpetrators actively worked to create this isolation. They limited the victims’ social interactions to increase dependence before taking advantage of them.

Scale and Nature of the Problem:

Before examining specific fraud techniques, it is important to establish why the architecture of the Unified Payments Interface (UPI) itself creates a structurally suitable target for fraud. Understanding these design characteristics provides the foundation for explaining how offenders exploit the system. The simple and easy-to-use nature of UPI makes it an appealing option for online banking: the user’s phone number is linked to their bank accounts, debit cards, and credit cards through a platform, which allows for instant transactions at no cost. Singh and Agarwal (2025) state that while this attribute provides convenience, it also means that a single phone number becomes the single point of compromise that gives the attacker access to all of the user’s financial information simultaneously.

Sharma, Matharu and Sinha (2026) surveyed 183 UPI users in the Shimla district. As the study was confined to a single district and employed a relatively small sample, its findings have limited generalisability to the wider Indian population. The study found that although 96.2% of respondents reported being aware of UPI fraud, 52.5% had nevertheless become victims. The authors concluded that awareness alone may not translate into safe behaviour, highlighting behavioural factors such as overconfidence and susceptibility to social engineering. By contrast, Singh and Agarwal (2025) surveyed 280 respondents in Varanasi, excluded 30 inconsistent responses, and analysed 250 valid cases. Using the Friedman ranking test, the study identified awareness (mean rank = 6.72) and education (mean rank = 6.42) as the most effective countermeasures against UPI fraud. The two studies are not necessarily incompatible. Rather, when taken together, they suggest that awareness and education are seen as the best defences, but they do not always prevent fraud in practice. The difference in district, sample size, and research design may also contribute to the contrast.

In a comparative study of India, Sweden, and the United States of America, it was found that losses due to card fraud in India grew at an average annual rate of 12.19% between 2011 and 2016 (Iyer, 2017). Sweden, however, despite having the highest percentage of card transactions among the countries studied, demonstrated the lowest average annual fraud growth rate of 4.83%, which is nearly a third of the growth rate in India (Iyer, 2017). According to the study, India’s higher growth rate could be a result of card payment systems being introduced relatively recently, when compared to Sweden and the USA.

Furthermore, Sweden’s digital payment ecosystem functions within a complex regulatory framework managed by Finansinspektionen, the Swedish Financial Supervisory Authority. This agency sets security rules, checks compliance, and inspects payment platforms. Digital payment providers must follow the Payment Services Act, anti-money laundering laws, counter-terrorist financing regulations, the General Data Protection Regulation, and the Payment Card Industry Data Security Standard. They are also required to implement strong authentication, encryption, tokenization, and transaction monitoring systems as necessary security measures (Abraham et al., 2023).

According to Mukhopadhyay and Mukhopadhyay (2024), despite becoming victims, a large proportion of people were unwilling to leave UPI as a means of payment. Only 7.4% of fraud victims switched entirely to cash, 8.8% reduced their use of digital payments, and 51.5% reported no reduction in UPI usage. The Pearson correlation coefficient between fraud victimisation and reduced UPI usage is -0.408, a moderate negative relationship that confirms victims are largely unwilling to exit the platform. Holtfreter et al. (2008) demonstrated that a 1-unit increase in routine consumer activity, such as purchasing something online, translates into a 61% increase in the odds of being targeted for fraud. The data indicates that since victims themselves are substantially reluctant to use different means of payment, people will almost always be exposed to and within the environment in which motivated offenders operate.

It is difficult to determine the true scale of the problem due to issues regarding under-reporting. Mukhopadhyay and Mukhopadhyay (2024) found that 20.6% of respondents who had experienced UPI fraud did not report the incident to the concerned authorities. This finding is based on a survey of 70 valid respondents conducted over a five-day period (20–25 April 2023). Sirajutheen and Abirami (2026) found that, during the “Golden Hour” (the critical period immediately following a cyber fraud during which prompt reporting maximizes the likelihood of freezing fraudulent transactions and recovering funds) of a fraud incident, 40% of fraud victims would incorrectly call their bank’s IVR (Interactive Voice Response) helpline instead of the National Cyber Crime Helpline (1930). Meanwhile, post-fraud grievance redressal mechanisms prove to be largely ineffective, with 33% of people finding them unsatisfactory and only 6% describing the bank’s response as very satisfactory (Sirajutheen and Abirami, 2026). While under-reporting is an issue in and of itself, the low recovery rate of assets further highlights the shortcomings of guardianship. According to Sharma and Singh (2024), ₹2294 crore ($274.14 million as per the 2024 exchange rates) was lost in cyber fraud, but only ₹57 crore ($6.81 million as per the 2024 exchange rates) was successfully recovered, showing a recovery rate of 2.5%. The low recovery rate indicates that the existing institutional mechanisms for recovering funds after UPI fraud are objectively ineffective. Such limited recovery could reduce the effectiveness of formal reporting systems and suggests that current guardianship mechanisms provide insufficient protection to victims.

How Fraudsters Operate and Why People Become Victims

One of the most prominent modalities of UPI fraud is social engineering (the psychological manipulation of the target’s behavior). Sharma and Singh (2024) present social engineering attacks in which fraudsters impersonate trusted authorities and create a sense of urgency regarding issues such as KYC compliance, blocked accounts, or failed transactions. Similarly, Wadkar et al. (2025) studied two UPI fraud cases involving educated and technologically aware victims who were deceived by messages warning of imminent electricity disconnection and directing them to malicious links. These studies demonstrate that digital payment fraud takes advantage of psychological pressure and trust rather than being a result of mere technological ignorance.

Other prevalent techniques include phishing and vishing (voice phishing), fake QR code scams, fraudulent payment request schemes in which fraudsters send payment requests disguised as payment receipts, and SIM swap attacks that redirect incoming OTPs to the fraudster’s device (Sharma and Singh, 2024; Singh and Agarwal, 2025). Sirajutheen and Abirami (2026) report that while awareness of phishing (67%) and vishing (70%) is relatively high among UPI users, awareness of more complex schemes (such as mule account operations, in which victims unknowingly launder money by renting their bank accounts to fraudsters) is relatively low, with only 6% of respondents being familiar with the term.

Singh and Agarwal (2025) point out that the platform connects a debit card, credit card, and bank account to one phone number. This means that one successful manipulation gives the offender access to several financial tools at once. Sharma and Singh (2024) also mention that digitization has not only widened the market for fraud but has also made criminals more skilled. Offenders keep adapting their techniques to keep up with the changes in the platforms they target.

Survey data from Sharma, Matharu, and Sinha (2026), drawn from 183 UPI users in Shimla district, indicates that 68.3% of fraud victims are graduates or postgraduates, challenging the assumption that higher educational attainment inherently protects individuals from digital financial fraud. As established above, 96.2% of the same respondents reported being aware of UPI fraud, yet 52.5% had nevertheless experienced victimization (Sharma et al., 2026). Singh and Agarwal (2025), meanwhile, identified fraud awareness as the highest-ranked preventive measure among surveyed stakeholders, assigning it a mean rank of 6.72. Taken together, these findings expose a structural gap: awareness and education are the most widely recommended individual-level defences against digital payment fraud, yet their presence has failed to prevent victimization at scale. Sirajutheen and Abirami (2026) conducted a correlation analysis that revealed a strong positive relationship between the frequency of digital usage and victimization (r = .825). This finding aligns with the Routine Activity Theory, which suggests that increased routine exposure creates greater opportunities for motivated offenders. Holtfreter et al. (2008), in a study of 922 adults, established a crucial distinction between the determinants of fraud targeting and the determinants of fraud victimization. Their logistic regression models demonstrated that routine consumer activity predicted whether individuals were selected as targets, producing a 61% increase in targeting odds per unit increase in activity. Individual traits, however, did not affect targeting at all.

Holtfreter et al. (2008) established that while RAT explains targeting, individual factors determine whether a targeted person becomes a victim. In the UPI context, fraudsters systematically exploit cognitive biases.

Several cognitive and social biases contribute to vulnerability to fraud. Overconfidence bias leads users to underestimate fraud risk and disregard security protocols and updates. FOMO (Fear of Missing Out) is exploited through fabricated urgency, such as claims of limited offers or imminent account suspension, to push users toward hasty decisions. Reciprocity bias, along with sympathy exploitation, is described as common in romance and distress scams. Confirmation bias, anchoring, authority bias, and groupthink are identified as social influences that interact with these factors to further increase susceptibility to manipulation and fraud (Sharma, Matharu, & Sinha, 2026).[1]  This reveals that exploiting the architecture of human decision-making has a vital role in achieving UPI fraud.

The comparison with DeLiema’s study is instructive here. In elder fraud, cognitive vulnerability arose from neurological decline, which is a form of diminished capacity that reduced the victim’s ability to evaluate threats accurately (DeLiema, 2018). In UPI fraud, cognitive vulnerability is manufactured when fraudsters design scenarios that activate the biases listed above, effectively inducing a temporary incapacity in otherwise competent individuals. The complex nature of the attack is what converts the targeted individual into a victim, regardless of education and awareness.

Victims and non-victims attribute fraud causation differently. Victims emphasize financial illiteracy and technological complexity, while non-victims emphasize cybersecurity vulnerabilities and slow investigation. This divergence suggests that victimization is associated with a shift in how individuals understand the threat (Singh and Agarwal, 2025).

Systematic Failure

The institutional response to UPI fraud victimization has largely been a refusal to accept responsibility. Wadkar et al. (2025) document that when two fraud victims reached out to a pseudonymous Bank ABC, the bank maintained that the transactions were authorized because the victims had shared their UPI PIN or one-time password (OTP). This was the case regardless of whether the victims did so under deceptive circumstances, leading the bank to conclude that the transactions were non-refundable. The RBI’s own framework reinforces this position, specifying that if a transaction occurs as a result of the customer sharing their credentials, the loss rests with the customer until the bank is formally notified (Wadkar et al., 2025).

Sharma and Singh (2024) state that social engineering attacks trick customers into disclosing information by impersonating legitimate authorities and manufacturing crises. However, the victim does not share credentials negligently. Rather, they are deceived into doing so. Yet the legal framework treats engineered deception and careless disclosure as equivalent, and holds the victim equally liable in both cases.

One of the most prominent legal gaps in India’s UPI consumer protection framework is the APP (Authorized Push Payment) loophole. According to Agrawal (2026), as per RBI[1] ’s 2017 Consumer Protection Circular, banks are obligated to cover losses from unauthorized transactions if reported within three days. APP fraud, however, operates by deceiving the victim into authorizing the transfer themselves, so the transaction is technically classified as authorized and falls entirely outside the protection framework.

The failure to provide an appropriate institutional response is evident in the recovery statistics. In 2022, the recovery rate was determined to be 2.5% (Sharma and Singh, 2024). This statistic highlights a significant failure of systemic guardianship. Furthermore, it was revealed that only 6% of respondents rated their bank’s grievance redressal process as “very satisfactory,” while 33% found it “unsatisfactory.” These perceptions align with the recovery outcomes, indicating that the current system is not effectively providing reliable protection  Sirajutheen and Abirami, 2026, p.301).

The 20.6% non-reporting rate among UPI fraud victims is not clear evidence of a lack of action. The authors suggest two possible explanations for non-reporting: that the grievance redressal mechanism may be difficult for victims to access, or that victims may lack sufficient awareness of how and where to report UPI fraud (Mukhopadhyay and Mukhopadhyay, 2024).

Furthermore, in elder fraud cases, capable guardians (financial institutions, social services, family members) consistently failed to identify victimization until losses had already occurred, at which point intervention was too late (DeLiema, 2018). When we look at UPI, the transactions are much more rapid, and the window for intervention is narrower. Therefore, the convenience of quick transactions also brings forth the issue of timely prevention by guardians.

Recommendations

The most significant reform that Indian policymakers can make is to close the APP liability loophole. ​​The United Kingdom introduced a mandatory reimbursement of up to £85,000 (approximately $113,000 as per the 2026 exchange rates) for APP fraud, requiring payment service providers to compensate eligible victims and strengthening incentives for fraud prevention (Agrarwal, 2026). A comparable reform in the Indian context would require banks and payment platforms to compensate victims of social engineering fraud, removing the current presumption that authorized transfer equals voluntary consent.

Legal frameworks governing digital payments must evolve continuously to match the pace of technological change (Ballaji, 2024). Amending the Information Technology Act (2000) and the Payment Settlement Systems Act (2007) is recommended to address evolving forms of digital payment fraud, support a fraud management framework, and implement AI for fraud detection (Banerjee et. al, 2025).

It has been established that a majority of users aware of digital payment fraud have become victims (Sharma, Matharu, and Sinha, 2026). Therefore, policymakers should design interventions that reduce the probability of a person falling victim to fraud at the moment of the transaction. An example of such could be a form of QR authentication before the approval of a transaction.

Furthermore, since 40% of users were unaware of the National Cybercrime Helpline (1930) during a fraud incident (Sirajutheen and Abirami, 2026), a simple display of the helpline in any UPI platform (such as Google Pay) would allow for timely action and early intervention.

Additionally, while many studies support using AI in fraud detection, there is little guidance on how to implement these systems effectively. Future research should focus on building and testing practical AI-based fraud detection frameworks that can be integrated into digital payment systems.

Several methodological limitations in the current literature also need to be addressed. Future research should use nationally representative samples rather than regional ones, as this would allow the findings to be generalized across India. Furthermore, future work should test which specific cognitive biases result in relatively greater victimization, so that prevention efforts can be prioritized. Finally, comparative research across a wider set of countries would help identify which regulatory and infrastructural conditions most effectively curb fraud growth.

Conclusion

This paper explains why India’s digital payment revolution driven by UPI has produced a growing class of fraud victims. The observations suggest that UPI fraud stems from structural factors that create opportunities for fraud, rather than solely from individual negligence. UPI’s architecture, in which a single phone number links a user’s bank accounts, debit cards, and credit cards, creates a single point of vulnerability, while its convenience ensures routine digital activity that places users in the realm of motivated offenders.

Routine activity governs who is targeted, but awareness and education alone do not determine who becomes a victim. Victimization depends on whether fraudsters succeed in manufacturing the cognitive conditions, such as urgency, authority, and overconfidence, that override otherwise informed judgment.

One of the most critical issues is the absence of effective guardianship. Recovery rates for stolen funds remain minimal, and redressal mechanisms are widely regarded as unsatisfactory. Additionally, the Authorized Push Payment loophole allows banks to treat engineered deception and careless disclosure as equivalent, leaving a large category of victims without reimbursement.

Therefore, closing the gap between UPI’s convenience and its vulnerability will require structural changes, not merely individual reforms. Tackling the APP loophole, updating the legal frameworks governing digital payments, addressing methodological limitations, and introducing safeguards at the point of transaction are of utmost importance. As UPI continues to expand as the backbone of India’s digital economy, the strength of its guardianship mechanisms is likely to determine whether this expansion will occur without a proportionate rise in victimization.

References

1. Abraham, Eva, et al. “Exploring the Regulatory Demands and Evolution of Payment Security Regulations for Digital Payment Platforms in Sweden.” Law and Economy, vol. 2, no. 10, 2023, pp. 18–26, https://doi.org/10.56397/LE.2023.10.03.
2. Agrawal, Sristy. “Cyber Fraud in Digital Payments: A Global Study on UPI Scams and Online Financial Crimes.” Indian Journal of Law and Legal Research, vol. 8, no. 2, 2026, pp. 6991–7014.
3. Association of Certified Fraud Examiners. “Fraud 101: What Is Fraud?” ACFE, https://www.acfe.com/fraud-resources/fraud-101-what-is-fraud. Accessed 4 July 2026.
4. Ballaji, Nima. “Consumer Protection in the Era of Digital Payments: Legal Challenges and Solutions.” Beijing Law Review, vol. 15, 2024, pp. 1268–1290, https://doi.org/10.4236/blr.2024.153076.
5. Banerjee, Adhip, et al. “Digital Payments in India: Exploring Emerging Issues in the Legal and Regulatory Framework.” Krytyka Prawa, vol. 17, no. 1, 2025, pp. 78–96, https://doi.org/10.7206/kp.2080-1084.753.
6. Cohen, Lawrence E., and Marcus Felson. “Social Change and Crime Rate Trends: A Routine Activity Approach.” American Sociological Review, vol. 44, no. 4, 1979, pp. 588–608. JSTOR, https://doi.org/10.2307/2094589.
7. DeLiema, Marguerite. “Elder Fraud and Financial Exploitation: Application of Routine Activity Theory.” The Gerontologist, vol. 58, no. 4, 2018, pp. 706–718, https://doi.org/10.1093/geront/gnw258.
8. Holtfreter, Kristy, et al. “Low Self-Control, Routine Activities, and Fraud Victimization.” Criminology, vol. 46, no. 1, 2008, pp. 189–220.
9. Iyer, Swarnalakshmi M. A Case Study on Monetary Fraud in a Cashless Economy. 2017. Purdue University, Master’s thesis. Purdue e-Pubs, https://docs.lib.purdue.edu/open_access_theses/1289.
10. Ministry of Finance, Government of India. Coordinated Efforts of Government, RBI and NPCI Accelerate Growth in Digital Payments. Press Information Bureau, 2025, Press Release ID 2240723, https://www.pib.gov.in/PressReleasePage.aspx?PRID=2240723&reg=3&lang=1.
11. Ministry of Finance, Government of India. UPI Completes 10 Glorious Years, Emerges as World’s Largest Real-Time Payments Platform, Anchoring India’s Digital Economy. Press Information Bureau, 2026, Press Release ID 2257087, https://www.pib.gov.in/PressReleasePage.aspx?PRID=2257087&reg=3&lang=2.
12. Mukhopadhyay, Nibedita, and Mahadeb Mukhopadhyay. “UPI Frauds: A Study on UPI Usage, Awareness and Impact in India.” International Journal of Research in Commerce and Management Studies, vol. 6, no. 6, 2024, pp. 179–197, https://doi.org/10.38193/IJRCMS.2024.6616.
13. Sharma, Ashish, and Yogender Singh. “Cyber Frauds in India’s Digital Payment Ecosystem: Risk, Impacts, and Regulatory Responses.” Educational Administration: Theory and Practice, vol. 30, no. 5, 2024, pp. 15326–15332, https://doi.org/10.53555/kuey.v30i5.8951.
14. Sharma, Reetika, et al. “Socio-demographic and Behavioral Determinants of UPI Fraud Vulnerability: A Descriptive Study from Shimla District, Himachal Pradesh.” Journal of Forensic Science and Research, vol. 10, no. 1, 2026, pp. 9–14, https://doi.org/10.29328/journal.jfsr.1001110.
15. Singh, Amit Kumar, and Krishna Kumar Agarwal. “An Overview of Digital Payment Frauds: Causes, Consequences, and Countermeasures.” Journal of Informatics Education and Research, vol. 5, no. 1, 2025, pp. 2297–2311.
16. Sirajutheen, T., and V. Abirami. “A Study on Analysis of Financial Fraud in the Indian Banking Sector.” International Advanced Research Journal in Science, Engineering and Technology, vol. 13, no. 3, 2026, pp. 297–302, https://doi.org/10.17148/IARJSET.2026.13347.
17. Vinoth Kumar, S., and Dharshan S. “An Analytical Study of Financial Frauds in E-Payment Systems in India.” International Journal of Creative Research Thoughts, vol. 14, no. 3, 2026, pp. d326–d331.
18. Wadkar, Prashant, et al. “Cybersecurity Challenges in Digital Payments: A UPI Fraud Case Study from India.” International Journal of Science, Architecture, Technology, and Environment, vol. 2, no. 10, 2025, pp. 227–232, https://doi.org/10.63680/ijsate1025030.024.

About the Author

Aadhavan Singh

Aadhavan is currently a Grade 12 student at La Martiniere College, India. He is a national-level shooter and passionate trekker.

The post Exploring How India’s Digital Payment Revolution Created a New Class of Fraud Victims: An Analysis of UPI Scams appeared first on Exploratio Journal.

Abstract

This paper focuses specifically on terrorism in India, with particular reference to Kashmir insurgency and the Naxalite (Maoist) movement, examining the structural, political and socio-economic factors that contributed to the emergence of these insurgent groups. 

The study is further situated within two principal theoretical lenses identified by Jeff Goodwin: the traditional perspective, which conceptualizes terrorism as a “weapon of the weak”, and the radicalization perspective. In addition, the paper critically reviews the work of prominent scholars whose analysis highlights diverse motivational factors. 

Building on this framework, it also identifies and analyses the principal causes underlying the Kashmir insurgency, highlighting political marginalization, cross-border influence, identity-based grievances, the legacy of the unresolved 1947 Partition dispute, and the impact of Hindutva politics have shaped the trajectory of conflict in the region. 

Extending this analysis, the paper examines the Naxalite Movement (Maoist) movement, tracing its roots to socio-economic inequalities, agrarian exploitation and the persistent marginalization of tribal communities, alongside the ideological influence of contested claims regarding external (particularly Chinese) involvement.

However, in reference to the Kashmir insurgency and the Naxalite (Maoist) movement, Terrorism in India cannot be narrowly confined to religious extremism alone. Rather, it must be understood as the outcome of intersecting structural, ideological and political factors. India’s complex geopolitical location and its history entrenched conflicts have further vulnerabilities in the internal security framework. In light of these findings, it’s important for the enforcement of inclusive governance, credible institutions and sustained socio-economic development to address root causes that perpetuate conflict and instability instead of militarization. 

1. Introduction

What are the causes and consequences of terrorism in India with reference to the Kashmir insurgency and Naxalite movement today? India, like many other countries around the world, has suffered several terrorist attacks over the last two decades. In this research paper, I aim to explore some of the factors that contribute to the rise of terrorism, and to examine its consequences on domestic politics and foreign relations. 

Terrorism has emerged as one of the most persistent and complex challenges in contemporary states, particularly in developing democracies such as India. Several studies identify socio-economic disparities, political exclusion, regional imbalances and perceived injustice as certain factors contributing to the emergence of terrorism. In this model, terrorism is often conceptualized not merely as indiscriminate violence, but as a strategic instrument employed by non-state actors.

Two prominent theoretical perspectives can help in contextualizing the phenomenon. The traditional view portrays terrorism as the “weapon of the weak”, suggesting that downtrodden groups resort to asymmetric violence and in contrast, the radical perspective interprets terrorisms as part of a wider process of escalating political violence, where ideological mobilization, identity-based politics and structured inequalities over the time.

This study points out two major challenges that have continued to shape India’s territorial security landscape: the Kashmir conflict and the Naxalite (left wing extremist) movement. The Kashmir issue represents a multifaceted conflict rooted in historical disputes, regional aspirations and cross-border interactions. Motivations which link to terrorism in this region include separatist aspirations, the influence of global jihadist movement, ideological divergences, perceived political marginalization, etc.

Arising from the 1960s in Naxalbari village of West Bengal, the Naxalite movement reflects a different but also an important form of political violence. Emerging from a radical communist ideology, the movement initially mobilized marginalized rural populations around issues of land redistribution, class exploitation and state neglect

By examining these two cases, this paper seeks to address our central research question: What are the principal causes of terrorism in India? Through a critical analysis of structural conditions, ideological motivations and state responses, the study aims to contribute to an understanding of terrorism as a product of intersecting political, economic and identity-centric factors. This study examines these two cases and offers policy recommendations. 

2. Defining Terrorism

It is important first to define the concept of “terrorism” before addressing my main research question. Terrorism is a highly contested term, and there is no common argument among the scholars on how to define it. However, the purposes of this paper, is to define terrorism in line with Goodwin (2019): “any tactic or set of tactics used by any government, group, organization, or individual, in pursuit of a political goal (broadly defined), which is intended to kill or harm civilians or noncombatants (as opposed to soldiers or political leaders) so as to frighten, intimidate, demoralize, provoke, or pressure other civilians and/or political leaders.” In short, this definition prioritizes that terrorism is primarily about the harming of civilians in order to intimidate others for political ends.

3. LITERATURE REVIEW: CAUSES OF TERRORISM

The section reviews literature on the causes of terrorism, highlighting two major arguments: one links to individual factors like religious fundamentalism and psychological traits while other traces to political marginalization and socio-economic inequalities and ethnic divisions. 

3.1 Individual-Level Explanations: Ideology and Psychology

To start with the first argument, older literature tends to emphasize that terrorists- were motivated by ideology, psychological predispositions, or religious fundamentalism (especially Islamic fundamentalism) (Pape 2003; Kramer 1990; Merari 1990; Post 1990). 

However, Pape (2003) argues that modern day terrorist attacks have not just limited to religious fundamentalists or to isolated men with certain psychological variations, but he gives the example of the LTTE (Liberation Tigers of Tamil Eelam), a group which recruited individuals who were Hindu Tamils but who were inclined towards Marxist ideology.

3.2 Structural and Political Explanations

The second type of argument focuses on some social grievances for example: political – and not religious, or individual psychological – motives of the terrorists. It has highlighted the motives i.e., to achieve specific political motives, change the policies in favour of them, to get financial backing (Crenshaw 2012; Abrahms 2008; Goodwin 2019). Crenshaw (2012) argues that violent acts are intentional political and strategic choices, rather than the inevitable result of social or psychological factors. In her view, terrorism functions as a “weapon of the weak,” used by actors who lack conventional means to influence government policy. Groups such as Euskadi ta Askatasuna (ETA) in Spain and the Provisional Irish Republican Army (PIRA) in Northern Ireland have used violence to pursue ethnic and religious rights Abrahms (2008) similarly argues that terrorism is the last and only resort for some groups to achieve their political goals.

Although the act of suicide terrorism might be irrational on an individual level. Thomas Schelling (1996) may have called it “the rationality of irrationality” – that is, the individual dies but the group is able to strategically demonstrate to the audience that it is serious about its political goals (Pape 2003). Terrorist groups are more likely to use suicide terrorism when there is a possibility of changing public opinion in favour them (i.e., in democracies). In the words of Boix and Rosato (2001): “The target state of every modern suicide campaign has been a democracy. The United States, France, Israel, India, Sri Lanka, Turkey, and Russia were all democracies when they were attacked by suicide terrorist campaigns, even though the last three became democracies more recently than the others.” Crenshaw (2012) similarly argues that terrorists do not just target communist regimes, but also target democracies.

Terrorism is closely linked to social and political conditions that create exclusions and injustice, which can be explained by some contextual examples on socio-political grievances. For example, when a minority group feels isolated in their own homeland, it forces them to retaliate or harm the civilians of their own nation or others. To make it clearer, we can take the example of attacks on Muslim community in India whether in Kashmir, Ayodhya or Godra Kand, which scorn the wounds of the neighbouring states which led to 26/11.  In short, historical rifts are also the causes which never cure the wounds of the nation. 

According to Edward Newman (2006) “These conditions-such as poverty, demographic factors, social inequality and exclusion, dispossession, and political grievances can be either permissive or direct.” This shows how terrorism is directly linked with environmental factors which degenerates the equality, rights and dignity of a human being. Ongoing social inequality and exclusion contribute to perception of injustice, while political grievances and dispossession weaken trust in state institutions. 

Furthermore, Terrorism is closely linked with the failures of human securities which can be understood as a consequence of insecurity and modernization where marginalized individuals resort to violence due to lack of opportunities and social breakdown for which O’Neill (2002) says, “human insecurity, broadly understood, provides the enabling conditions for terrorism to flourish”. As a result, insecurity of a human leads to its action which might harm the community. Also, terrorism is considered as the last resort to accomplish their needs and desires when peaceful methods don’t pay off. To understand with a few examples: Afghanistan, Sudan, Algeria, Yemen, Georgia, Pakistan. Somalia- which are the fundamental core of the operations conducted by the groups all over the world because of poverty.

In accordance with earlier explanations, this perspective offers a different account of the causes of terrorism. As per Martha Crenshaw (2012), she firstly makes a point on Modernization which creates a link between the individuals via easy transportations and communications.   Urbanization is one of the best tools which comes under modernization, as it increases the number of mobility and accessibility of targets and methods, which she termed as “urban guerilla warfare”, which was seen in Latin America in the 1960s.

This approach indicates that terrorism is not just driven by ideology or political grievances but is also shaped by structural transformations in society. Processes such as modernization and urbanization create environments that increase connectivity, mobility and the concentration of targets, thereby making terrorist strategies more feasible. 

4. THEORETICAL FRAMEWORK

Theoretical framework explain terrorism through two perspectives- first is Traditional theory, finds terrorism as “the product of weakness and/or desperation of some rebel or state (a ‘weapon of the weak’)” when there is a lack of capacity in pressurizing their opponent by their non-violent or conventional acts. The second, The Radicalization Perspective, which believes that “not all radicals may be terrorists, but all terrorists are radicals”. Radicalization is understood as “a process leading towards increased use of political violence”. Both theories ultimately points to a common goal i.e. Political motives. (Jeff Goodwin, 2019). Together, these perspectives indicate that terrorism is not merely a tactical choice but also a political constructed strategy shaped by both opportunities. 

5. INDIAN TERRORISM LANDSCAPES

This section focuses on the insurgencies faced by India.  India is a secular and multicultural nation. Since 1947, conflicts with Pakistan have led to wars. The Kashmir dispute emerged after the 1947 Instrument of Accession, leading to prolonged conflict between India and Pakistan. But India is not just facing such insurgencies by Pakistan, internal insurgencies continue which harm the peace and harmony of the nation. Apart from Kashmir related issues there is another big issue, i.e. The Naxalite or Maoist movement began in 1969 in Naxalbari and spread across states.

5.1 On Kashmir Conflict

Historical Background

The Kashmir dispute emerged during the 1947 partition of British India. Princely states were permitted to join either country. Jammu and Kashmir, ruled by Maharaja Hari Singh, had a Muslim-majority population but was governed by a Hindu ruler who initially attempted to remain independent. In October 1947, tribal fighters from Pakistan entered the region, prompting the Maharaja to seek military assistance from India, leading to the first war between India and Pakistan. This conflict was referred to the United Nations, which called for a ceasefire and proposed a plebiscite to allow Kashmir to determine their future. Unfortunately, the plebiscite was never conducted and the region was divided. In Curfewed Night, Basharat Peer presents this history as lived trauma. The memoir briefly outlines the unfulfilled promise of self-determination. Rather than analysing diplomatic negotiations, the book highlights how unresolved political conflict translated into militarization, curfews and insecurities in the Kashmir Valley. Having outlined the historical foundations of the dispute, the discussion now moves to the factors that have shaped the rise and continuity of militant violence in Kashmir. 

Drivers of the Kashmir insurgency

Many possible motives can be drawn for attacks in India, which is provided by K. Alan. Kronstadt (2008, pg. 6), according to some reports, radical Islamic sentiments play a crucial role, as radicals wanted a certain revenge for their Indian Muslim community in Ayodhya, Uttar Pradesh- 1992 and at Godra, Gujarat- 2002. And it is known that the radicals were inspired by Al-Queda’s brand of global jihadi ideology.  Some views indicate LeT has aimed for Kashmir’s separatism as their primary goal.  The Kashmir problem is related to the state’s claim by which state, which was very clearly demarcated since 1947. By military, Line of Control separated Muslim majority Jammu and Kashmir from Pakistan which controlled Azad Kashmir. Secretary Rice has speculated that the goal of the attackers was “probably to stir up trouble between Pakistan and India.” 

Terrorism violence in India cannot be explained by a single ideology; it results from the interaction of religious radicalization, geopolitical rivalries and unresolved territorial disputes. The Kashmir conflict and India-Pakistan tensions provide a strategic environment where extremist groups try to internationalize the issue and escalation between the two states. Thus, terrorism also functions as a strategic tool to influence regional politics.

In line with Shivaji Khemnar (2018), there are ethno-nationalist terrorism, religious terrorism, left wing terrorism, Narco terrorism. Roots of massive terrorism in India, are various in ideology, poverty, regional imbalance, strong worship about religion. India suffers attacks from Bangladesh, Nepal and Pakistan. 

Following this, Vani. K. Borooah (December 2008), provided some factual data of the attacks held in Kashmir. Over the period of 1998-2004, 784 terrorist incidents in India which resulted in 3008 deaths, and 61% of them and 55% of casualties occurred in Jammu and Kashmir alone. Apart from native-born terrorism, Kashmir conflict has attracted Al-Qaeda who have viewed this as a large part of Islamic strategy. A pamphlet titled “Why are we waging jihad” by Lashkar-e-Taiba includes the spread Islam in every corner of India. (Haleen 2004; Borooah 2008). Islamist groups were responsible for 126 attacks out of which 12 were suicidal attacks, out of which 6 were by Lashkar-e-Taiba, 2 were by Al-Mansurian (LeT) and 2 were by Jaish-e-Mohammad.

This is important statistical evidence which shows the concentration of terrorist incidents in J&K, the analysis largely focuses on quantitative data and military group involvement. Hence, the data highlights the scale and external linkages of terrorism. 

The Kashmir revolt is traced from the middle class and their frustration as it lacks employment opportunities (Sumit Ganguly; Prem Shankar Jha). Why do democratic political systems tend to experience higher incidents of terrorist attacks? 

It is explained in the lines “The growth of political assertiveness of minority communities in democratic states is virtually inevitable. As minorities acquire increased literacy and education, they will become more conscious of their political rights and will seek to assert them. Nondemocratic, poly-ethnic states can suppress minority demands for political participation through co-optation, coercion, or repression.” (Sumit Ganguly, 1996). 

Hence, this argument can be interpreted as rising awareness and mobilization especially among educated minorities– can create pressures that political institutions struggle to manage. However, this explanation may overemphasize political participation as a cause of terrorism, while underestimating other factors such as regional, external support to militant groups and state responses. Thus, democratic openness may create space for dissent, but it does not necessarily make democratic the only and primary cause for terrorism. 

Terrorism cannot be explained solely in terms of social and political factors; rather politics plays a crucial role in shaping and concealing issues that are deliberately prevented from being revealed. In Kashmir there are 3-way perspectival divisions: Hindus in Jammu, Buddhist in Ladakh and Muslims in Kashmir and it has created a paradigm that Kashmir is Muslim and is Anti-National. It gives a reflection of the Kashmiris struggle “against India’s brutalities”, where they were illegally detained, tortured, imprisoned and exploited. 

Under the governance of NDA, ideology of Hindutva was followed and spread in Kashmir without considering Muslims in Kashmir, which has resulted in the rise of attack frequency.

Hindutva-driven political strategies have indirectly increased tensions in Kashmir by making many Muslims feel excluded and unheard. Debates around Article 370 and 35A, along with efforts to fully integrate the region have added to this feeling of insecurity. At the same time, the growing support between Hindu nationalist groups and some Kashmiri Pandits has deepened divisions. These tensions and grievances are used by militant groups, which keep the cycle of conflict and violence going. (Mridu Rai, 2019)

Her focus is on political narratives, identity politics, and ideological agendas- particularly debates around Article 370 and 35A- which has shaped perspectives of marginalization in Kashmir. However, this argument primarily emphasizes the role of Bhartiya Janta Party and Hindutva politics. Nevertheless, political factors play a significant role, since politics shapes governance and national decisions-making which remains crucial in understanding terrorism and instability. 

The BJP’s Kashmir Policy: Continuity and Change from Vajpayee to Modi

While the Kashmir conflict is often explained through historical, social and security-related factors, political dynamics also constitute a significant dimension, as political parties frequently shape policies and narratives in alignment with their ideological orientations and strategic interests. 

In line with Mridu Rai (2019) there are 2 eras of the BJP government, which depicts the conditions in the Kashmir valley since the 1990s. According to the report, since 1990 Kashmir has experienced less governance in the conventional sense and more population control enforced through fear and violence. The passage further argues that the Indian state under the Modi government has intensified this security-oriented approach, framing it within the ideological vision of the Hindu Rashtra.

It presents a comparative analysis of the BJP under Atal Bihari Vajpayee and under Narendra Modi, particularly in relation to Kashmir and Pakistan. It suggests that Vajpayee’s tenure represented a relatively peaceful phase in India’s Kashmir policy. Vajpayee included this approach in the slogan “Insaaniyat, Jamhooriyat aur Kashmiriyat”, signalling a framework attached with humanity, democratic process and recognition of Kashmir identity.

In contrast, the Modi government has described as adopting a more assertive and militarised way toward Pakistan. Public celebration of “surgical strikes” and repeated warnings of retaliations show increased aggression. This external posture is said to be mirrored internally in Kashmir, where stronger security measures have been justified in the name of countering cross-border terrorism. 

 It argues that organisationally, the BJP has undergone significant transformation since 2014. While centralized under Vajpayee, but more monolithic under Modi’s leadership, with reduced space for internal dissent and increased subordination of regional leadership to central command. Since 2014, the author suggests, incidents of hostility and violence against such groups have been more visible. 

Cross-border Militancy in Jammu & Kashmir

Numerous attacks in Kashmir are observed, like Mumbai Train Blast (2006), Mumbai Attack 26/11(2008), Mumbai Bombing (2011), Pathankot Airbase Attack (2016), Uri Attack (2016), Pulwama Attack (2019), Bengaluru Attack (2023), Pahalgam Attack (2025). 

The most recent Pahalgam attack (22nd April 2025), which resulted in the death of 27 security personnel and injury to over 40 civilians (Adil, Sohail & Farid, 2025).  Pahalgam was a core of tourism along with the transit point for Annual Amarnath Yatra pilgrimage, which was then considered as the attack on for both its scale and its symbolism. By the Indian Ministry of Home Affairs, Pakistan’s military group Lashkar-e-Taiba was accused of this insurgency. (Tanveer, 2025). This time India retaliated with fearlessness by its operation named “Operation Sindoor” on 7 May 2025.

The paragraph concludes with the call of, that Kashmir has been called by Arundhati Roy, a ‘real theatre of unspeakable violence and moral corrosion.’ (Muhammad Feyyaz 2019). Geopolitics plays a significant role with all events happening in and around the world. Geography is what stabilizes or destabilizes a nation’s security and development. Pakistan’s desire to cover Kashmir in its territory is not just in the interest of Islamic fundamentalism, instead, its geographical, hydrological and strategic factors. It is important for Pakistan to maintain a powerful position in Kashmir, essentially for water security and advantageous position against India.  

Repercussions of Armed insurgency

Since 1947, this prolonged conflict in Jammu and Kashmir has had wide- ranging impacts beyond security. It has led to environmental damage, economic decline (majorly tourism), has disrupted education and regular life, along with serious mental health issues as per Dr. Mushaq Margoob, Professor & Head, Post Graduate Department of Psychiatry, Government Medical College, Srinagar. Taken together, these have weakened the region’s social and economic stability. Overall, terrorism has deeply affected Kashmir’s development and the life-style of the people. 

5.2 THE NAXALITE (Maoist) MOVEMENT

Historical Background and Origins

After examining the conflict in Kashmir, it is equally important to turn to another major internal security challenge persisting in India–the Naxalite Movement. The Naxalite movement, also known as Maoist insurgency in India, originated in 1967 in the village of Naxalbari in West Bengal as a radical peasant uprising against those land inequalities and exploitative agrarian structures. Inspired by Maoist ideology and the principles of the armed revolution, the movement sought to challenge what is known as a semi-feudal and oppressive state system. The movement was ideologically influenced by revolutionary doctrine linked with Mao Zedong and political development in China during the 1960s. However, this was largely evolved as a domestically rooted response to socio-economic inequalities, tribal marginalization and governance deficit within India. Initially driven by demands for land redistribution and social justice for marginalized communities–which then gradually expanded into the central and eastern parts of Indian states, forming what is often referred to as the “Red Corridor”. Over the period of time, the movement evolved from localized agrarian resistance into an organized armed insurgency. Manohar Parrikar Institute for Defence studies and Analysis (2010) states that the Naxalite movement was driven by radical Marxist- Leninist ideology which led to the formation of the CPI (M-L). Multiple internal challenges seen in: the Central-Eastern area (Red Corridor), North-East India by groups like NSCN & ULFA. (Aditya Dasgupta, Kishor Gawande and Devesh Kapoor; 2017)

Understanding the structural and immediate causes that helps in the emergence and persistence of this movement becomes essential to comprehend its relevance in India’s internal security landscapes. 

Structural and Socio-Economic Determinants

In 2008, Prime Minister-Manmohan Singh described Naxalism as “the greatest threat to our internal security”. The pace of the Naxalite movement suggests not merely state incapacity, but deeper structural contradiction within India’s developmental model, has been the driving force to such insurgent acts. While the Indian state pursued parliamentary democracy and economic modernization, large parts of rural and tribal populations remained excluded from land reforms, political representation, and access to justice. 

In line with European Foundation for South Asian Studies Publications: Naxalbari, the village that gave its name to the movement, was the site of the peasant revolt, started by communist leaders against owners of the state. The Naxalbari uprising was not a mere reason for agrarian anger but an ideological rupture within Indian communism. It marked a rejection of parliamentary gradualism and signalled the re-emergence of revolutionary violence as a legitimate political instrument among sections of the left. Members of Communist Party of India (Maoist), notably Charu Majumdar, Kanu Sanyal and Jangal Santhal who initiated this movement. Drawing inspiration from Maoist ideology, these leaders advocated a strategy of long armed struggle aimed at overthrowing the existing socio-economic order through agrarian revolution. The uprising reflected a vision articulated by the so-called “Siliguri Group”, which called for a militant path to political transformation. The formation of CPI in 1925 institutionalized communist ideology in India drawing inspiration from the global Marxist-Leninist movements and adapting them to the colonial Indian context. In its early years, the CPI was committed to a Marxist-inspired mass revolution but later faced internal and external pressures that reshaped its strategy. In 1951, after consultations with Stalin- who advised participation in democratic politics rather than armed revolt–the CPI reconsidered its revolutionary path. The ideological reshaping laid the groundwork for subsequent fractions within the left movement ultimately which contributed to the emergence of radical Maoist. Disillusionment with the CPI’s gradual and pro-soviet orientation- particularly its commitment to parliamentary democracy and perceived abandonment of revolutionary militancy–led to the break away at the Seventh Party Congress in Calcutta. The CPI-M thus emerged as a distinct political formation that rejected what it characterized as soviet “revisionism” and aimed to reaffirm a more orthodox and militant Marxist line, drawing ideological inspiration from the Chinese Communist model. 

Naxalite emerged as a radical response to deep rooted socio-economic inequalities. Its primary causes according to Sanjay Seth (1995), were as follows: The very first is, Structural Contradictions of Capitalist Development- Capitalist expansion generates contradictions between socialized productionand private appropriation leading to crises which intensify exploitation and weaken legitimacy, creating conditions for resistance.Second is seen as the Formation and Organization of the Proletariat- capitalist modernization concentrates workers in the factories and cities, fostering class solidarity and collective resistance and also creates a basis for revolutionary mobilization. Moving forward, Marxist Reorientation towards Peasant Agency in ‘Backward’ Societies- Marxist revolutionary theory shifts emphasis to peasant agency in backward societies, viewing peasant struggles against feudalism as a key driver for revolutionary change.Along with that Gap between Consciousness and Marxist Political Objectives- Peasant resistance, shaped by local and traditional identities, which differ from Marxist class ideology, creating a gap that leads to ideological tensions. In addition to the initial factors, Misrecognition of Peasant Subjectively by Communist Parties- where communist parties misread peasant consciousness, projecting revolutionary intent, leading to a strategic and ideological disconnect. And last but not the least, Instrumentalization of Peasant Struggles by Vanguard Parties- communist parties often used peasant struggles for their own goals, directing movements beyond peasants’ intent, leading to dissatisfaction.

While Sanjay Seth highlights contradictions between peasant consciousness and Marxist political objectives, the Naxalite movement demonstrates how ideology and lived experience interact dynamically rather than mechanically. Peasant resistance in India was not purely “false consciousness” nor fully revolutionary; rather, it evolved through local grievances–land alignment, state repression and caste hierarchies–that were gradually reframed within Maoist political language. Thus, the movement’s radicalization can be understood as a process of ideological translation.  

It follows that the persistence of Naxalite exposes the gap between India’s procedural democracy and substantive socio-economic justice, particularly in marginalized regions such as Chhattisgarh and Jharkhand. Governance failures, extractive development and weak reforms lead the insurgent appeal, framing it as internal security issues. The growth of Naxalism reflects deeper structural inequalities in India’s development model rather than only a security issue.              

Integrated approaches to mitigate Naxalite insurgency

Although the Naxalite insurgency has emerged from the multiple structural and socio-economic causes discussed above, the Indian government has also introduced several strategies to mitigate the movement. In line with Manohar Parrikar Institute for Defence Studies and Analyses (April, 2010) The Government of India, has adopted a two-pronged strategy, to address Naxalism, combining security measures and governance interventions.The government set up an Empowered Group of Ministers, a Coordination Centre, and a special task force to improve coordination. It focused on a Law-and-Order Approach through police modernization and long-term deployment of centre forces, with the centre bearing costs (around  ₹1,100 crore) to support states. Another approach the government adopted is Social Integration Approach, focusing on development through schemes like the Backward District initiative and BRGF, allocating funds to affected areas. It also implemented state-level support and infrastructure programs like PMGSY to improve living conditions and reduce Naxal influence. 

Sino- Maoist connections

After examining the internal causes and state responses to Naxalite insurgency, it also becomes important to assess whether external factors contribute to these insurgent activities. This analysis helps determine whether policy responses should extend beyond domestic measures to include broader strategic and foreign policy considerations. 

There are two principal stands of argument, one contends that China has no involvement with Naxalite or Maoist movement in India, while the other claims that China bears responsibility for supporting or facilitating these insurgencies.

First argument notes, Maoist groups’ linkages with militant, Northeast insurgent, and radical Islamist organizations for logistical support pose a significant internal security challenge to India, while multiple factors suggest Chinese involvement in sustaining Left-Wing Extremism (LWE). An analysis by the Vivekanand International Foundation quotes The Home Secretary, G.K.Pillai in November 2009, where he was confident with the supply chains of arms by China to Maoists in India, “Chinese are big smugglers and suppliers of small arms. I am sure that the Maoists also get them.”

The Home Minister, P. Chidambaram, acknowledged that Naxalites procure weapons through cross-border routes via Nepal, Myanmar, and Bangladesh, but expressed uncertainty regarding any direct assistance from Chinese state actors. Nevertheless, repeatedly Chinese made arms and communication equipment from Maoist cadres are often cited as indicative of possible linkages between Maoist groups and China. 

To line RSN Singh (2010), the Home Secretary has asserted that China is a major supplier of small arms and suggested that Maoist Groups may procure weapons of Chinese origin. In 2004, a significant supply of light and medium machine guns and ammunition were seized at Chittagong port in Bangladesh, reportedly sourced from China. Bangladesh intelligence officials further alleged that leaders of United Liberation Front of Asom and National Socialist Council of Nagaland (Isak-Muivah) visited Kunming to obtain arms, some of which were allegedly routed to Maoist groups operating in Nepal and India. 

Second, it argues that China has nothing to do with any insurgency, according to the statement issued by the Ministry of External Affairs (May 17, 2012) that India raised concerns with China regarding reported support for anti-India armed groups, prompting China to assure a policy of non-interference and deny such support. China has officially denied supporting anti-government forces and emphasized its policy of non-interference; the dialogue mechanism indicates that India remains cautious and will continue to monitor dimensions of internal threats.

6. Conclusion

In conclusion, the study demonstrates that terrorism is a complex and evolving phenomenon that cannot be understood through a single framework. The discussion of various scholarly definitions, theoretical perspectives and the examination of cases such as Kashmir and Naxalite illustrate that terrorism emerges from the interaction of ideological, political, socio-economic, and historical factors, earlier, terrorism was often narrowly associated with religious extremism; however, contemporary conflicts increasingly reveal an overlap between terrorism and civil wars. This growing convergence has created distinguish terrorism from other forms of political violence, consequently, the absence of a universally accepted and precise definition of terrorism continues to challenge the international community in formulating consistent responses.

From the Indian perspective, geography plays a dual role. While India’s geographical position has historically facilitated cultural exchange, economic interaction and strategic connectivity. It has also exposed the country to security vulnerabilities and cross-border tension. In addition to these historical and geographical factors, ideological influences have also shaped certain forms of insurgency in India. The influence of socialism and communalism, particularly in regions characterized by socio-economic disparities and governance gaps, has contributed to the emergence of structural grievances, including poverty, land alienation and the neglect of marginalized communities. 

The findings of this research suggest that addressing terrorism requires a multidimensional approach. While security measures remain important, long term solutions must also focus on ensuring inclusive development for all sections of society. Future research should therefore explore innovative perspectives that move beyond purely historical explanations and instead examine how structural reforms, effective governance and socio-economic justice can help mitigate the conditions that sustain violent movements. 

Although the complete eradication of terrorism may be unrealistic, meaningful progress can still be achieved through stronger international cooperation, stricter accountability for states that have terrorist networks, and institutional reforms. Finally, the persistence of terrorism highlights not only individual responsibility but also government failures. When the government is unable or unwilling to address citizens’ aspirations and grievances, extremist ideologies find space to grow. In this context, terrorism may be defined as “a strategic and radical practice aimed at influencing state policy through coercion or violence.” 

References

  The Oxford Handbook of Terrorism. (2019). United Kingdom: OUP Oxford.

  Golder, B., & Williams, G. (2004). What is “Terrorism”? Problems of Legal Definition (SSRN Scholarly Paper No. 1351612). Social Science Research Network.

Pape, R. A. (2003). The Strategic Logic of Suicide Terrorism. American Political Science Review, 97(3), 343–361.

Chapter- THE LOGIC OF TERRORISM, Terrorist behaviour as a product of strategic choice by MARTHA CRENSHAW

Terrorism Studies: A Reader. (2012). United Kingdom: Routledge.

 Abrahms, M. (2008). What terrorists really want: Terrorist motives and counterterrorism strategy. International Security, 32(4), 78-105 

Kronstadt, K. A. (2008). Terrorist attacks in Mumbai, India, and implications for US interests (No. CRSR40087).

 Khemnar, S. (2018). Causes and effects of terrorism in India: An overview. International Journal of Applied Information Systems, 12(15), 29-32.

 Panda, J., & Pankaj, E. (2025). Proxy Wars and Silent Partners: The Pahalgam Attack.

Gaibulloev, K., & Sandler, T. (2008). The impact of terrorism and conflicts on growth in Asia, 1970-2004 (No. 113). ADBI Discussion Paper.

  Ullah, A., Qingxiang, Y., Ali, Z., & Anees, M. (2018). Terrorism in India as a Determinant of Terrorism in Pakistan. Asian Journal of Criminology, 13(1), 57-77.

  Saeed, K., Khan, N. A., Arif, I., ur Rehman, A., Rehman, M., & Maheen, N. (2025). FROM PAHALGAM TO CEASEFIRE: THE 2025 INDIA–PAKISTAN CONFLICT, ITS HISTORICAL ROOTS, AND DONALD TRUMP’S MEDIATION ROLE. Pakistan Journal of Social Science Review, 4(4), 1173-1204.

 Ganguly, Š. (1996). Explaining the Kashmir insurgency: political mobilization and institutional decay. International Security, 21(2), 76-107.

Mukherjee, S. (2018). Colonial origins of Maoist insurgency in India: Historical institutions and civil war. Journal of Conflict Resolution, 62(10), 2232-2274.

Rai, M. (2019). Kashmiris in the Hindu Rashtra. Majoritarian state, 259-280.

 Seth, S. (1995). Interpreting revolutionary excess: The Naxalite movement in India, 1967–1971. Modern Asian Studies, 29(1), 205–229. https://doi.org/10.1017/S0026749X00016123 

Singh, R. S. N. (2013). Is China waging a proxy war through the Maoists? Indian Defence Review, 28(2).

Peer, B. (2010). Curfewed night: A frontline memoir of life, love and war in Kashmir.

 https://www.un.org/en/global-issues/countering-terrorism

  https://www.efsas.org/publications/study-papers/economy-and-ecology/

 https://www.wider.unu.edu/publication/prolonged-effects-terrorism

https://www.researchgate.net/profile/Sanjay-Seth/publication/249879533_Interpreting_Revolutionary_Excess_The_Naxalite_Movement_in_India_1967-71/links/5f7b476992851c14bcaf0ddb/Interpreting-Revolutionary-Excess-The-Naxalite-Movement-in-India-1967-71.pdf

Britannica, T. Editors of Encyclopaedia. (n.d.). Naxalite. In Encyclopaedia Britannica. Retrieved [Date], from https://www.britannica.com/topic/Naxalite

Srikakulam peasant uprising. (n.d.). In Wikipedia. Retrieved [Date], from https://en.wikipedia.org/wiki/Srikakulam_peasant_uprising

South Asia Citizens Web. (n.d.). The Naxalite movement: Background and developments. Retrieved [Date], from https://www.sacw.net/article1622.html

International Socialism Journal. (n.d.). The ironies of Indian Maoism. Retrieved [Date], from https://isj.org.uk/the-ironies-of-indian-maoism/

 Political Science Institute. (n.d.). Naxalbari uprising: Genesis of Naxalite movement in India. Retrieved [Date], from https://polsci.institute/india-democracy-development/naxalbari-uprising-genesis-naxalite-movement-india/

https://www.vifindia.org/article/2011/march/25/China-Naxalite-linkages-Gauging-its-dimensions

https://www.mea.gov.in/articles-in-indian-media.htm?dtl/19721/Q598+China+Aiding+Insurgency+in+North+Eastern+States

Source: MP-IDSA https://share.google/ZTQOuCwQstdTYhuLy

Source: EFSAS https://share.google/U55MzH7Zm8dnOYYlH

Source: Vivekananda International Foundation (VIF) https://share.google/LpfQkGBjOJUAXfhgx

ACKNOWLEDGEMENT

I would like to express their sincere gratitude to Dr. Mashail Malik, Assistant Professor, Harvard University for her invaluable guidance, intellectual support, and constant encouragement throughout this research. I also thank her for the insightful discussions and helpful feedback on the initial drafts of this manuscript.

I am grateful to the Department of Political Science at Banasthali Vidyapith for providing the necessary institutional resources to conduct this study.

On a personal note, I wish to thank my parents and friends for their unwavering moral support, patience, and encouragement during the course of this work. Their belief in my efforts was a constant source of motivation.

Finally, I thank the anonymous reviewers for their constructive comments and suggestions, which significantly contributed to improving the quality of this paper.

About the author

Disha Tyagi

Disha Tyagi is currently pursuing a Master’s degree in Political Science from Banasthali Vidyapith. Her academic interests revolve around International Relations, South Asian regionalism, geopolitics, diplomacy, public policy, global governance, trade relations, conflict studies, security studies, and international cooperation. Among these, terrorism and regional instability remain one of her strongest areas of interest and research. She is particularly interested in studying how terrorism impacts regional integration and development, especially in areas such as Kashmir and Left-Wing Extremism in India.

Her academic journey has been enriched through active participation in seminars, conferences, webinars, debates, and international academic engagements. She attended the BRICS International School: New Generation in March 2026 and participated in the Global Studies Program organized by the Russian Federal University in February 2026. She also attended the seminar on “WTO MC14 Outcomes: Future of Multilateralism and Implications on India’s Trade Agenda” held on 5 May 2026, which deepened her understanding of trade diplomacy and multilateral institutions. Additionally, she participated in webinars such as “The Structural Causes of World Poverty” by Thomas Pogge on 23 April 2026 and “India’s BRICS Presidency 2026” by Antara Ghosal Singh on 8 May 2026. These academic interactions enhanced her understanding of global justice, development politics, and emerging geopolitical transformations.

Beyond academics, she has actively participated in Model United Nations and parliamentary simulations, including the Youth Parliament (UNGA Portfolio) held during SANSAD 2025 in Delhi. She has also completed online courses including “Successful Negotiation: Essential Strategies and Skills” by the University of Michigan, “Learning How to Learn” by the University of Arizona, and “Writing in the Sciences” from Stanford University, along with an offline communication course at the British Council, Delhi.

The post From Global Causes to Local Realities: Terrorism in India with Reference to Kashmir and the Naxalite Movement appeared first on Exploratio Journal.

Abstract

This paper aims to address how second chances within duty and redemption tie together in the context of medieval knighthood and honor in the poem Sir Gawain and the Green Knight. The poem Sir Gawain and the Green Knight is a poem about second chances and redemption, telling the moral that while mistakes are inevitable, second chances allow individuals to achieve redemption in their duty. Through critical analysis of Sir Gawain and the Green Knight and the character Gawain the Glorious, Gawain’s character is shown to be one who constantly makes mistakes throughout his journeys. Gawain is offered second chances through the form of Lord Bertelak’s game, where his honesty is tested as a knight. Even when Gawain inevitably fails, he is offered partial redemption in the form of a nick he receives on his neck from the Green Knight. This portrayal is significant of a character seen as a model knight. The message in Sir Gawain and the Green Knight uses a previously glorified Gawain to send the message that, despite the expectation of perfection within duty, continuous redemption and improvement through second chances from past mistakes is honest to one’s duty.

In a profession as prized as medieval knighthood, the concept of life, duty, mistakes, and how they interconnect has been a long-debated concept. The relationship between one’s life and one’s duty has been a long-standing question that has been challenged since the beginning of literature. A significant example of this is the famous poem Sir Gawain and the Green Knight, written during the late 14th century, where the relationship between duty and life is explored through the character Gawain, a knight of King Arthur.

Throughout the entirety of Sir Gawain and the Green Knight, Gawain’s role as a regarded knight is continuously tested throughout his adventures. As a knight of King Arthur, he is prized within Camelot. Further, he is a relative of King Arthur himself and is portrayed as a model knight (Jobe). However, within the poem and Gawain’s journey, his mistakes in his duty are ultimately what lead him to survive and redeem himself to live up to his prized identity. Throughout the poem, multiple factors like Sir Gawain’s identity as a knight, the antagonist Green Knight, and Lord Bertalak’s game rest on the underlying themes of second chances and mistakes. Sir Gawain and the Green Knight is a poem about redemption within duty. While mistakes within duty are inevitably made, second chances still frequently present themselves to give opportunities to redeem oneself in their duty while also holding consideration of their life.

The first instance of a mistake leading to a second chance is in the opening scene in Camelot during the feasts of Yule. When the Green Knight approaches the court with a challenge, not a single knight rises to meet him, despite chivalry and bravery being included in a knight’s honor code. This includes Gawain. It is only when Arthur is forced to arise and almost meets the Green Knight’s challenge that Gawain stands up and takes his challenge:

“Gawain bv Guenevere

Toward the king doth now incline:

‘I beseech, before all here,

That this melee may be mine.”

…

“I am the weakest, well I know, and of wit feeblest;

And the loss of my life would be least of any;

That I have you for uncle is my only praise;

My body, but for your blood, is barren of \mrth;

And for that this folly befits not a king,

And ’tis I that have asked it, it ought to be mine,

And if my claim be not comely let all this court judge,

in sight.” (ll. 338-361, Borroff)

In this scene, Gawain stands up and declares, “I beseech before all here that this melee may be mine…if my claim be not comely let all this court judge in sight.” Not only does Gawain claim responsibility as the player in the Green Knight’s game in front of the court, but he also asks the entire court to hold him responsible for his word, which shows he’s committed to seeing the challenge through. Further, in his words against himself, “I am the weakest…and wit feeblest…the loss of my life would be least of any…” In this, he acknowledges his imperfections as a knight and his willingness to sacrifice himself for the court and King Arthur, something that nobody else had stood up for. Despite Gawain’s delayed acceptance of the Green Knight’s challenge, a lapse in his duty to be brave and chivalrous, he swiftly makes it up by taking up and committing to the responsibility of the Green Knight’s challenge from King Arthur’s hands. This shows redemption in his duty and the rest of the court of being an honorable knight despite his initial mistake because he stood up despite nobody else doing so.

Gawain’s next mistake is shortly after his redemption when he takes the responsibility of the Green Knight’s challenge. In his game with the Green Knight, where they exchange a blow for a blow with the Green Knight’s axe, Gawain deliberately makes his cut a fatal one. He was prompted by King Arthur to be intelligent about his blow, as well as tempted by the Green Knight with the outcome of a fatal blow. The Green Knight says: “and if I spend no speech, you shall speed the better: You can feast with your friends, nor further trace” (ll. 409-410). The Green Knight’s words are deliberately tempting — he brings up Gawain’s friends as well as the possibility of him feasting, if he didn’t have to receive a cut in return. Gawain falls for the bait, and by attempting to make it so the Green Knight would not be able to give him a blow back in return, he does not follow the traits of a medieval knight. It does not display bravery or justice to try to slip out of a game Gawain volunteered for, more especially after his poignant declarations to the entire court of his willingness to lose his life. However, the Green Knight survives, and Gawain is also expected to receive the same blow in return as per the rules of the Green Knight’s game. Thus, despite Gawain’s attempts to go through a loophole in the Green Knight’s game being a mistake, it prompts the rest of the journeys, and, at the time for Gawain, unknown second chances for him to redeem himself as a knight as well as his life. This is also explored in Georges Gusdorf’s “The Game of Chance: Moral Ambiguity in Sir Gawain and the Green Knight.” The loophole reflects a larger ethical ambiguity, as Gawain effectively trades his honor as a knight for a better chance to keep his life. Both preservation of life and duty are important values in Sir Gawain and the Green Knight, and the loophole that Gawain utilizes shows the complexity that accompanies abiding by both duty and life (Gusdorf). The poem later reflects:

Now take care, Sir Gawain,

That your courage wax not cold

When you must turn again

To your enterprise foretold. (ll. 487-490, Borroff)

In this scene, the speaker directly addresses Gawain with a declaration that he needs to be courageous in the future. This reads similar to a warning that Gawain has journeys ahead of him that require his courage as a knight, the very thing the game was testing and relying on. It serves as an acknowledgment that Gawain’s future journey is not so simple, because if the journey was just him receiving the cut, he would not need to ensure that his “courage wax not cold.” Further in the story after the year has passed and Gawain’s expected meeting with the Green Knight approaches, he reflects on his future journey.

“There was much secret sorrow suffered that day

That one so good as Gawain must go in such wise

To bear a bitter blow, and his bright sword

lay by.

He said, “Why should I tarry?”

And smiled with tranquil eye;

“In destinies sad or merry,

True men can but try.” (ll. 558-565, Borroff)

This scene gives the reader some insight into Gawain’s emotions and thoughts regarding his journey and his perceived imminent death. He rhetorically asks himself why he should delay his presumed imminent death, then follows it up by saying “True men can but try” when faced with destiny. While it isn’t clear exactly what Gawain is referring to, based on his future destiny to die, and his pondering over why he should delay the inevitable, it can be assumed Gawain is referring to living past his destiny of death. This shows that despite his acceptance of his imminent death, there is still a part of him that wishes to try and live and redeem himself to be a “true man” who would try and live. This offers the reader some insight on the complexities of Gawain’s current emotions, showing that he feels “tranquil.” Despite his duty as a knight and his expected sacrifice, there is still a part of him that wishes to live and have a second chance at life — something that Gawain says a “true man” would do. This implies that Gawain feels that, to be true to himself, he must also try to live. This displays Gawain’s uncertainty in his future journey despite it being something he is bound by duty to since he took up the Green Knight’s challenge, along with consideration of his own life and his hesitancy about death. However, it also shows a development in Gawain’s attitude toward life and duty, as he also considers the preservation of his life despite an imminent death. This is also explored in Marjorie Nicolson’s “Second Chances and Self-Knowledge in Sir Gawain and the Green Knight.” Gawain’s consideration of his imminent death as a result of a mistake shows a sense of maturity, as he still opts to try as a “true man” would, displaying adherence to staying true to his duty despite his situation. Despite his past mistakes, he resolves to move forward with consideration of his duty (Nicolson). Later, when on his journey, he faces a near-death predicament before his duty is fulfilled. The poem reads:

Thus in peril and pain and predicaments dire

He rides across country till Christmas Eve, our knight.

And at that holy tide

He prays with all his might

That Mary may be his guide

Till a dwelling comes in sight. (ll. 733-739, Borroff)

This scene is another example of a second chance at living appearing for Gawain. After trekking to the point where he was on death’s door, he prayed with vigor for a chance to get guided to his destination. His prayers were answered, and soon after, he discovered Lord Bertilak’s court. It’s important that Gawain was on death’s door and was specifically given a second chance because it tells us that not only does he want to live to fulfill his duty, but he was also offered a second chance to do so through his well-timed discovery of Lord Bertilak’s court. This is once again another second chance for Gawain, but also specifically in the context of his necessity to survive to fulfill his duty, providing Gawain an opportunity to fulfill his duty despite his past mistakes when hesitating to take up the Green Knight’s challenge. However, it’s also important to consider whether or not Gawain’s second chance in this case was divine intervention from his prayer to Mary, or a second chance he redeemed from his own actions. But despite the way his life was saved being uncertain, Gawain was still presented with a second chance — whether it was one he seized from his own redemption or one he prayed for, representing a second chance nonetheless. These instances of second chances and Gawain’s uncertainty about his duty and life show the complexities the poem portrays about second chances in duty. The structure of the poem also reflects these repeated trials where Gawain faces second chances, as also analysed in Derek Pearsall’s “The Structure of Sir Gawain and the Green Knight.” The four fit structure shows a section where Gawain is tested once again to face a choice he must make, until finally, in the fourth fit, he’s set to face all of his previous choices and ultimately given a second chance to live, reflecting the structural support of each second chance. (Pearsall)

The next second chance that Gawain takes is when he arrives at the Lord’s court, in the third fit. When initially meeting the Lord, the Lord proposes a game of exchange to Gawain. The poem reads:

“And Gawain,” said the good host, “agree now to this:

Whatever I win in the woods I will give you at eve,

And all you have earned, you must offer to me;

Swear now, sweet friend, to swap as I say,

Whether hands in the end be empty or better.

“By God,” said Sir Gawain. “I grant it fortwith!

If you find the game good, I shall gladly take part.” (ll. 1105 – 1110, Borroff)

This shows the Lord’s offer of a second chance, despite Gawain’s initial ignorance of the true nature of the Lord’s game. However, the Lord’s game adds complexity because it depends on one Gawain’s duty as a knight, which a lack of brought him to his death — he needs to be honest to the Lord. This is a second chance to redeem his duty to better himself from his initial mistake that led Gawain to what he perceived as his imminent death. Gawain, however, successfully engages in the exchange of winnings, taking this second chance for his life and utilizing it. The only exception in this game was the last day, when he did not give the Lord the girdle out of the desire to save his life. The girdle’s role in the games is also significant. In Jane H.M Taylor’s “The Girdle and Its Meaning in Sir Gawain and the Green Knight”, the girdle is shown to be a badge of humility, stemming from a desire to live, but eventually evolves into a partial redemption after Gawain’s mistakes and through his second chances (Taylor). Gawain’s intentions with the girdle and its representation, and Gawain’s honesty regarding the Lord’s game is an interesting juxtaposition between Gawain’s honesty and his mistakes — while he was honest and redeemed himself in his duty by being honest, the mistake he made when he was dishonest stemmed from Gawain’s desire to live. On the third day of the exchange of winnings, when Gawain is considering lying to save his life, the poem reads:

When he gains the Green Chapel to get his reward:

Could he escape unscathed, the scheme were noble!

Then he bore with her words and withstood them no more,

And she repeated her petition and pleaded anew,

And he granted it, and gladly she gave him the belt,

And besought him for her sake to conceal it well. (ll. 1857 – 1862, Borroff)

In this next passage, we can see Gawain deliberately trying to be dishonest and “conceal” what would keep him alive. While the action of Gawain preserving his life is arguably ingrained in human nature, his dishonesty still is not representative of the values that a knight upholds — a mistake in the context of his duty. This mistake is later revisited when Gawain faces the consequences of the Lord’s game. When Gawain came to the Green Chapel to face the Green Knight, who was revealed to be the Lord, he was told he would receive one hit with the axe due to lying on the third day of the Lord’s game. While receiving the stroke, however, Gawain flinches and is criticized by the Green Knight. The poem reads:

“Bestow but one stroke, and I shall stand still,”…

And his [Gawain’s] shoulders shrank a little from the sharp iron.

Abruptly the brawny man breaks off the stroke,

And then reproved with proud words that prince among knights.

“You are not Gawain the glorious,” the green man said,

“That never fell back on field in the face of the foe,

And now you flee for fear, and have felt no harm:

Such news of that knight I never heard yet!

I moved not a muscle when you made to strike” (ll. 2267 – 2274, Borroff)

This passage shows the Green Knight is directly criticizing Gawain’s bravery and courage, which is part of his duty. He denies Gawain being himself, stating that Gawain is “not Gawain the glorious” who had “never fell back on the field in the face of foe.” This is significant, as the Green Knight denying Gawain being himself and “glorious” is him denying Gawain being a good knight who never feared “in the face of foe.” The Green Knight also criticizes Gawain for being a coward, saying “now you [Gawain] flee for fear.” Here, the Green Knight is denying Gawain’s knightly qualities, essentially saying by flinching, he hasn’t been a good knight. This is especially significant considering the honor accompanying medieval knighthood, as explored by Derek Brewer’s “The Ethics of Honor in Medieval Knighthood.” The honor and duty that accompany medieval knighthood were broken by Gawain’s fear of death, as Gawain’s frailty in the face of death highlights his imperfections as a knight. (Brewer) Further, Gawain also does make a mistake prior to his action of flinching because he lies about being still but flinches anyway, saying “I shall stand still” and yet “his shoulders shrank a little from the sharp iron,” representing him making a claim of being brave and not holding himself to it — a mistake in his duty to be brave as a knight. The Green Knight feigns a second hit, in which Gawain doesn’t flinch, and on his third hit, he does not behead Gawain — only nicks him. The Green Knight explains his intentions of the hit, saying:

“I owed you a hit and

you have it; be happy therewith!

The rest of my rights here I freely resign.

Had I been a bit busier, a buffet, perhaps,

I could have dealt more directly, and done you some harm…

True men pay what they owe:

No danger then in sight.

You failed at the third throw,

So take my tap, sir knight.” (ll. 2341-2357, Borroff)

In this passage, after Gawain’s third hit is revealed to be because he lies, as the Green Knight says, “you failed at the third throw.” However, the Green Knight didn’t behead him, showing that one simple nick was enough for Gawain “pay[ing] what [he] owe[d]” and also referred to him as “true man,” implying Gawain’s honesty despite the nick being due to him lying. This implies that if he’d fully lied and been dishonest, he would have been hit thrice — one for a lie on each day — and likely killed. Gawain, in being honest in the game of winnings, had unknowingly redeemed himself from his mistake at the feast last year when attempting to cheat the game by being honest and coming to fulfill the game; he was properly fulfilling what he should be doing as a knight, other than his mistake of dishonesty, which he was punished for. Further, despite his punishment, Gawain was still given a second chance because he wasn’t killed — he still had his life to make up for the mistakes he made of lying. The Green Knight also considers the context of Gawain’s lie, saying:

“So is Gawain, in good faith, to other gay knights.

Yet you lacked, sir, a little in loyalty there,

But the cause was not cunning, nor courtship either,

But that you loved your own life; the less, then, to blame.” (ll. 2365-2368, Borroff)

Here, the Green Knight also directly addresses his dishonesty and says that, despite his dishonesty not being true to his duty as a knight, he can’t be blamed for just wanting to live. The incompatibility of these concepts is also explored in Jesse Roberts’s “Chivalric Duty and Human Fallibility”. Both the honors of knighthood and Gawain’s methodology of preserving his life and his own human fallibility force Gawain to address the tension between these two values. (Roberts) This represents a more nuanced view of Gawain’s lie, building on the portrayal of life and duty in the poem. Despite his duty, Gawain had prioritized his life over upholding the values of his duty. Even then, he is still constantly faced by the weight of his mistakes, but at the same time, given constant second chances to improve and redeem himself. The nick he got for his dishonesty is a final representation of that second chance, and how life ties into duty — he was given a final second chance to live and uphold his duty after learning from the several mistakes he made. Gawain addresses this, saying:

“Your cut taught me cowardice, care for my life,

And coveting came after, contrary both

To largesse and loyalty belonging to knights.” (ll. 2379 – 2381, Borroff)

Gawain directly says he learned more about the duty that belongs to knights by saying the Green Knight “taught [him] cowardice…and loyalty belonging to knights” telling us that his experience prior to the games was lacking but he learned from the entire ordeal with the Green Knight. This shows that his redemption did affect the way he viewed his duty as a knight, and now he has a better view of how to “Care for [his] life” along with harnessing the loyalty belonging to knights. This change in Gawain is also shown in the way the poem describes him. It reads:

Wild ways in the world our worthy knight rides

On Gringolet, that by grace had been granted his life.

He harbored often in houses, and often abroad,

And with many valiant adventures verily he met…

The nick on his neck he naked displayed

That he got in his disgrace at the Green Knight’s hands…

“For where a fault is made fast, it is fixed evermore…” (ll. 2479-2512, Borroff)

Gawain is directly described to be “a worthy knight,” which shows that he is now truly a knight following his duty, and directly addresses his mistake and how it is now “fixed evermore.” His opportunity for this was only possible because of the second chance to live from his duty he got from the exchange game, showing that the message portrayed in Gawain’s character is that redemption in duty is achieved through second chances despite mistakes and mishaps made throughout your life. He did not portray perfection in his duty, but instead had a balance between redemption in his duty and consequences — something that the nick on his neck represents. He did not redeem himself fully, but his partial redemption allowed him to learn from the past mistakes he had made within his duty.

Throughout Sir Gawain and the Green Knight, there is a repetitive and continuous pattern of making mistakes within one’s duty and redeeming oneself. Zooming out, the portrayal of the character Gawain — who is highly praised for being a compassionate and chivalrous knight — openly making mistakes, is unorthodox. These ideas of Gawain as a perfect knight are explored in Kennis Jobe’s “A Model Knight: Sir Gawain, Chivalric Contradictions, and Grief in Medieval Literature.” This different portrayal of him is significant, considering Gawain’s previous portrayal as a model knight in previous works featuring Gawain, such as L’arte Périlleux and Claris et Laris. (Jobe) However, the appeal in a story like this shows in its message of redeeming oneself within duty. It has a certain ignorance toward the concept of perfection at duty and embraces the priority of your life and humanity in the actions of constant improvement, despite past portrayal as a model knight. Ultimately, this poem makes a powerful statement on the complexities of the relationship between life and duty, illustrating a message that, despite the weight of expectations and the want of perfection, the honest way to be true to one’s duty is achieved through constant improvement from past mistakes and seizing second chances to be better.

Works Cited

Borroff, Marie, and Laura L. Howes. Sir Gawain and the Green Knight: An Authoritative Translation, Contexts, Criticism. W. W. Norton & Company, 2022.

Pearsall, Derek. “The Structure of Sir Gawain and the Green Knight.” Speculum, vol. 52, no. 4, 1977, pp. 619–640.

Gusdorf, Georges. “The Game of Chance: Moral Ambiguity in Sir Gawain and the Green Knight.” Modern Philology, vol. 83, no. 1, 1985, pp. 67–78.

Brewer, Derek. “The Ethics of Honor in Medieval Knighthood.” Journal of Medieval History, vol. 14, no. 1, 1988, pp. 39–48.

Jobe, Kennis. “A Model Knight: Sir Gawain, Chivalric Contradictions, and Grief in Medieval Literature.” 2023. Louisiana Tech University, Louisiana Tech Digital Commons, https://digitalcommons.latech.edu/cgi/viewcontent.cgi?article=1104&context=theses

Nicolson, Marjorie. “Second Chances and Self-Knowledge in Sir Gawain and the Green Knight.” Studies in Philology, vol. 82, no. 3, 1985, pp. 281–294.

Taylor, Jane H.M. “The Girdle and Its Meaning in Sir Gawain and the Green Knight.” The Review of English Studies, vol. 56, no. 225, 2005, pp. 362–374.

Roberts, Jesse. “Chivalric Duty and Human Fallibility.” Speculum, vol. 70, no. 4, 1995, pp. 797–820.

About the author

Ananya Kota

Ananya Kota is an 11th grader at Los Altos High School. She interns at her local newspaper as a reporter and is a community editor for her local student-run journalism publication. She is a peer tutor at her school and is the co-founder of a local youth lit-mag.

The post Second chances and Redemption in Sir Gawain and the Green Knight appeared first on Exploratio Journal.

Abstract

This study investigates how gender stereotypes are performed and transformed through viral TikTok trends. While prior research has documented the persistence of gender norms in traditional media, less is known about how they circulate in participatory digital platforms driven by algorithmic visibility. Using a qualitative content analysis of 150 viral TikTok videos across 15 gender-related trends (2023–2025), this project examined both behavioral scripts and identity labels—formats that exaggerate gendered behavior or crystallize personality types. Each post was coded for tone, creator gender, and orientation toward stereotypes, with audience engagement analyzed through comment sections.

Findings reveal four recurring patterns: exaggerated performance, ironic reinforcement, gendered authenticity, and audience co-production of meaning. Together, these show that TikTok functions not as a passive mirror of gender norms but as a participatory system that constantly remixes them. The study contributes to ongoing debates about humor, irony, and digital performance, showing how participatory media both challenge and sustain cultural stereotypes.

Introduction

On TikTok, small jokes can turn into shared language overnight. A creator jokes, “She’s a 10 but keeps five-minute voice memos for every minor inconvenience,” and thousands reply with their own versions. Another sound—“I’m just a girl”—sets off a wave of videos performing femininity with a wink, while the comments debate whether it’s satire or self-own. These aren’t just jokes; they’re the raw material for how people talk about gender.

TikTok has become one of the most powerful cultural stages of our time. Trends that begin as small jokes or offhand performances can spread to millions of people within days, repeating until they feel woven into everyday language. These trends don’t just spread quickly; they invent new ways of talking. Many lean on gender, packaging familiar stereotypes in new forms or inventing fresh labels that quickly spill outside the app and into everyday speech. These labels, along with repeated jokes about behavior, act as a script for how we see one another. What starts as humor begins to shape how people describe each other, how they understand themselves, and how gender is talked about in the culture at large.

This project asks: How do gender stereotypes appear in TikTok’s most viral trends, and what patterns emerge in how they are reinforced, challenged, or redefined across genders? Some of these patterns show up as recurring behaviors—repeated jokes about actions tied to gender. Others take shape as labels or characters, shaping how people talk about each other offline. Both matter. The behaviors show how everyday actions are framed through gender expectations, while labels and characters harden into recognizable identities that can stick to people in everyday conversation.

Key terms. In this paper, scripts refer to repeatable formats that exaggerate behavior (such as jokes or skits), while labels describe shorthand identities like “pick-me girl” or “sigma male.” Both circulate through trends; TikTok’s viral formats are built around shared sounds, captions, or hashtags. Together, they show how the platform transforms gender norms into repeatable and remixable content.

The research design gathered a set of TikTok’s most viral, gender-related trends and analyzed them collectively. By examining both the scripts that exaggerate everyday behaviors and the labels that solidify into identities, the project traced the gendered messages these trends convey, who produces them, and how they shift when applied to mixed-gender contexts. From this analysis, four main patterns emerged: exaggerated performance, ironic reinforcement, gendered authenticity, and audience co-production of meaning. Each reveals something distinct about how gender circulates through TikTok’s participatory culture. 

Exaggerated performance refers to the way creators dramatize gendered behaviors until they become caricatured. These performances make stereotypes visible but also contribute to their recirculation through repetition.

Ironic reinforcement describes the ambiguity that occurs when a creator mocks a stereotype while simultaneously repeating it. Because irony is difficult to detect on TikTok’s fast-moving feed, satire often looks very similar to sincerity, which allows stereotypes to spread even when users intend to critique them.

Gendered authenticity highlights how creators perform “realness” in different ways depending on gender. Women often express authenticity through openness, confession, or vulnerability, while men typically frame it through logic, control, or self-discipline. These differences show that even authenticity follows gendered scripts.

Audience co-production of meaning captures how comments, stitches, and duets shape the interpretation of a post. Meaning becomes collaborative rather than fixed, and audience participation keeps stereotypes active long after the original video is posted.

Together, they show that gender on TikTok is never static but constantly being performed, reinterpreted, and reshaped through humor and participation. The goal was not just to collect examples, but to see the larger dynamics at work: what fades as a passing joke, what hardens into a stereotype, and what gets flipped into something subversive. By recognizing those dynamics, this study aims to show how humor-driven digital culture shapes how gender is understood and talked about today.

Conceptualizing Gender Stereotypes in Media

The media has long served as one of the most influential sites for shaping how gender is understood. Classic theories of gender stereotyping emphasize how portrayals of men and women become cultural scripts that audiences internalize and reproduce (Eagly et al., 2019). These scripts are rarely neutral. Men are typically framed as decisive, assertive, independent, or unemotional, while women are cast as passive, nurturing, or even decorative (Hentschel et al., 2019). These scripts gain power from repetition. Over time, these portrayals appear natural and normalized, giving the impression that they reflect reality instead of actively shaping it.  

Research confirms that gender stereotypes remain persistent across media forms. Santoniccolo et al. (2023) show that mainstream representations continue to objectify and sexualize women, while Lauzen (2018) documents women’s ongoing underrepresentation both on screen and behind the camera on television. Studies of professional settings also reveal similar processes: job advertisements often contain language that subtly reinforces gendered expectations (Sczesny et al., 2024). Even scientific communication is not exempt from this; Chen (2024) finds that controversial science dialogue online often carries gendered assumptions. Together, these studies demonstrate that stereotypes remain embedded in communication, but that leaves a gap in how they circulate in participatory environments like TikTok, where ordinary users rather than producers drive representation. This project addresses that gap by tracking how gender stereotypes and scripts are recycled or reshaped in viral trends on participatory platforms like TikTok.

Humor, Memes & the Reinforcement of Stereotypes

Humor is one of the most enduring ways stereotypes are transmitted. Jokes that target gender may appear harmless, but they reinforce bias by presenting stereotypes as socially acceptable (Ford & Ferguson, 2004). Memes work similarly: they depend on instant recognition, which often means they repeat familiar gendered tropes (Meghana & Vijaya, 2020). Recent studies confirm that sexist memes can influence emotional and moral processing, making discriminative attitudes seem less problematic (Paciello et al., 2021). These findings show that humor and memes play a serious role in stabilizing gender norms. 

At the same time, humor has the potential to destabilize stereotypes. Exaggerating gendered behaviors to extremes can expose how unnatural those stereotypes really are.  (Matamoros-Fernandez, 2023) shows how features like duets, stitches, and sound reuse encourage endless remixing of jokes, sometimes reinforcing stereotypes and sometimes tearing them down. Meme analyses by Sultana (2025) and Mihailescu (2024) show that memes can be flipped to question or debate social roles. This tension, between reinforcing and challenging, is significant to TikTok humor, which dominates viral content. This project, therefore, examines not only how jokes reinforce gender stereotypes but also when and how they create cracks.

Digital Labeling in Online Culture

Online culture frequently condenses complex identities into shorthand labels. Nilsson (2024), for instance, documents how the “I’m just a girl” TikTok trend packaged gender performance into a recognizable cultural script, while Tanner (2025) analyzes “sigma male” discourse as a form of toxic communication that reframes masculinity. In both cases, labels crystallize stereotypes while also creating opportunities for new forms of identity work. 

Labels are rarely descriptive alone. They almost always will imply judgment. To be called a “pick-me girl” means being critiqued for seeking male validation, while “sigma male” often mixes parody with aspiration. Huber and Baena (2023) show that women scientists on TikTok sometimes embrace labeling strategies to increase their visibility, though it risks reinforcing old stereotypes about women in STEM. Steinke (2024) expands this, noting that female creators both embrace and resist labels as part of identity work. Yang (2023) conceptualizes this process more broadly, describing memes and labels as “cultural scripts” showing how memes and shorthand terms shape identities online. These examples show how online labeling both simplifies and amplifies identity work. On TikTok, those labels spread even faster because anyone can adopt, parody, or remix them. This project follows that insight by examining how viral TikTok labels work to reinforce, reshape, or challenge stereotypes.  

TikTok’s Amplification of Gendered Trends

What makes TikTok different from earlier forms of media is its speed and its algorithm. Its “For You Page” curates content in ways that don’t just reflect popularity but actively construct it (Varmazyar & Cardama, 2023). Dillon et al. (2023) found TikTok content often reproduces bias, while Yin and Abdullah (2024) found that negative portrayals of women spread quickly through trending content. In other words, stereotypes on TikTok are not just reproduced, but they are accelerated by design. 

At the same time, TikTok’s constant remixing changes how these trends play out. A single sound may generate thousands of versions. some repeating stereotypes, others mocking them. Remixing matters because it’s how meaning gets made on TikTok: each reuse of a sound or caption gives an old idea new life, sometimes as humor, sometimes as critique, and sometimes as both at once. Suarez-Garcia et al. (2024) demonstrate how sexism appears in multimodal video data. Matamoros-Fernandez and Farkas (2023) highlights how humor and amplification interact to make harmful content that becomes both visible and normalized. Yet platforms like TikTok can also enable resistance. Huber and Baena (2023) show how women scientists use TikTok to disrupt expectations and assert their own new forms of visibility. Taken together, this research positions TikTok as both an amplifier and medium for resistance, reflected by their spreading of stereotypes at scale, while also simultaneously offering space for rebuttals. This study builds on that work by examining how gendered scripts and labels circulate in viral trends and whether they are reinforced, parodied, or redefined.  

Data and Methods

TikTok as a Research Site

TikTok has become one of the fastest-growing social media platforms over the past five years, with estimates projecting its global user base to exceed 1.5 billion monthly active users by 2025 (Business of Apps, 2025; Boeker & Urman, 2022). Unlike traditional media, TikTok is participatory, meaning users don’t just consume content; they actively produce and remix it. TikTok’s duet and stitch features, tools that allow users to place their videos alongside or directly after another creator’s post, also make trends highly interactive, enabling users to respond, critique, or extend a trend in ways that shape how gender stereotypes circulate. This makes it a uniquely useful site for studying gender stereotypes. Prior scholarship has shown how stereotypes are embedded in traditional media like television, film, and advertising (Goffman, 1979; Lauzen, 2018), but TikTok provides an opportunity to observe how those scripts spread, change, and gain traction in real time. While TikTok’s algorithm is central to this process, it is not transparent and cannot be studied directly. Instead, this project examines its manifestations: the trends that receive disproportionate visibility and circulate widely on the platform. As a young woman familiar with TikTok culture, I approached this research both as a participant and observer. My interpretations were informed by an understanding of platform humor, but I aimed to stay reflexive about how my own perspective might shape what I noticed. Because TikTok’s algorithm determines visibility in opaque ways, examining trends themselves, rather than trying to access the system behind them, became the most effective way to study how gendered content gains traction on the platform.

Defining and Selecting Trends

On TikTok, a “trend” refers to a repeatable format that thousands of users adapt with their own variations, often built around a shared sound, hashtag, or caption (Ling et al., 2021). In this study, I focus on two types of trends. The first are scripts, which take the form of jokes or skits that exaggerate a behavior. The second are labels, short phrases that mark out an identity or archetype. These trends matter because they reduce gender stereotypes into recognizable forms that can easily be repeated and remixed, ensuring their circulation across diverse audiences. Each reuse or remix of a trend pushes the stereotype further, sometimes reinforcing it and sometimes twisting it into parody or critique (Meghana & Vijaya, 2020).

Sampling Strategy

For this project, I examined viral trends from 2023 to 2025. Because TikTok does not provide a fixed threshold for virality, I developed my own operational definition: a trend counts as “viral” if an individual post reaches at least 100,000 views or 10,000 likes. These thresholds were chosen because they indicate visibility beyond a creator’s immediate follower base and align with prior research that uses engagement ratios as markers of wide circulation (Zhou, 2024).

Trends were identified through two complementary strategies. First, I used an “influencer-first” discovery, beginning with well-known creators across niches. Second, I used trend-first discovery, searching TikTok’s sound pages and hashtag feeds to locate formats gaining traction across multiple creators. To determine whether a trend was spreading beyond a single creator network, I examined whether the same sound or hashtag appeared across diverse accounts, especially when creators were unconnected by follower networks or genre. Using both strategies allowed me to capture trends that spread top down through influencers as well as those that emerged more organically from everyday users.

Dataset and Coding

The dataset included 15 viral trends, each represented by approximately 10 posts, for a total of about 150 posts. These trends reflected both male-coded and female-coded stereotypes. Before selecting videos for analysis, I applied several exclusion criteria: videos in languages other than English, reposts or spam content, advertisements, and posts where either creator’s gender or target gender could not be reasonably inferred. These exclusions helped keep the sample consistent and ensured that all posts could be analyzed using the same categories.

The dataset included 15 viral trends, each represented by approximately 10 posts, totaling around 150 posts. These trends reflected both male-coded and female-coded stereotypes. Each post was logged in a Google Sheet, with one row per video and the following structured fields:

• Trend Name: The viral format or meme category
• URL: The direct link to the post, ensuring retrievability
• Creator Gender: The gender identity that the creator presented in the video or on their profile, recorded as male, female, both (for group accounts), or unclear
• Target Gender: The group toward whom the stereotype, label, or joke was directed, recorded as male, female, both, or unclear
• Tone: The manner in which the stereotype was delivered, categorized as serious, ironic, parody/satire, or ambiguous
• Valence: The overall orientation of the post toward the stereotype, recorded as reinforcing, countering, or mixed
• Notes: A free-text column where I recorded my own observations about each post
  

The “Notes” section included details that did not fit neatly into categories—such as how the creator used captions, body language, or sound; whether the performance exaggerated a stereotype for humor; or if the video seemed to blur critique and reinforcement at the same time. In addition to my own interpretation, I also noted patterns in audience response by scanning top comments and engagement. These reactions provided insight into how viewers were reading the video—whether they laughed along, pushed back, or reinterpreted it. Together, this combination of creator performance and audience reception gave a fuller picture of how gendered trends were being interpreted on TikTok.

Because this project was conducted by a single researcher, no formal inter-rater reliability test was possible. Instead, I used a reflexive coding process to strengthen consistency. After coding the first 30 posts, I revisited my original labels for tone, valence, and target gender, compared them with later entries, and refined the category definitions to reduce ambiguity. I then re-coded a subset of earlier videos using the updated definitions to ensure that my interpretations were stable over time. While this is not a substitute for multi-coder reliability, it provides a structured way to check for consistency in a single-researcher qualitative study.

Table 1. Descriptive Overview

The “hybrid” tone category included videos where creators shifted between sincerity and parody, or where tone could not be cleanly separated because irony and seriousness were blended within the same performance.

Table 2. Overview of the 15 TikTok Trends Included in the Dataset

Sample of Composition

The final dataset included 150 posts drawn from 15 distinct viral TikTok trends, with around ten posts collected for each. Together, these posts captured how gendered humor, labeling, and performance show up in TikTok’s most visible spaces between 2023 and 2025. As shown in Table 1, about half of the videos were created by female-presenting users, around a third by male-presenting users, and the rest by creators whose gender presentation was mixed, collaborative, or not clearly identifiable. This makes sense given TikTok’s overall user base and the dominance of feminine-coded humor, lifestyle, and commentary genres that tend to drive virality on the platform.

When looking at tone, roughly one-third of the videos were serious in how they portrayed gender roles, while about 27 percent used parody or satire. A smaller portion blended humor and sincerity—often flipping between the two in the same clip. This mix of tones captures something distinct about TikTok itself: the way sincerity and irony constantly overlap. The same sound, caption, or format can mean completely different things depending on who performs it and how. That’s part of what makes these trends so powerful—they invite both imitation and reinterpretation.

In terms of valence, nearly half of the videos reinforced familiar gender stereotypes, while others either pushed back against them (around 14 percent) or combined both reinforcement and critique (about 38 percent). The overlap between these categories is important. It suggests that even when stereotypes are being “mocked,” they still circulate widely and can easily be taken at face value. Many creators seemed aware of this tension and leaned into it, using humor or exaggeration to test how far they could go without fully endorsing what they were representing.

Finally, the Notes section helped fill in what the numbers couldn’t show. Here, I recorded details about performance style, creator intent, and audience reactions—especially the comment sections, where people often made the meaning of a post explicit. Sometimes viewers laughed along and reinforced the joke; other times, they called it out or reinterpreted it. These responses often shifted how a post could be read, showing how meaning on TikTok is a shared construction between creator and audience. Together, these observations point to a platform where gender norms are not just displayed but constantly negotiated, blurred, and remixed in real time.

Results

From the 150 TikTok videos analyzed, several key patterns emerged in how gender stereotypes were performed, shared, and interpreted. Before turning to those patterns, it helps to understand what these “trends” look like in practice. On TikTok, a trend is usually a short, repeatable format — often a joke, sound, or role-play — that thousands of users adapt with their own twist. Many of the gendered trends in this study use humor and exaggeration to explore how men and women are expected to behave. For instance, videos under “She’s a 10 but…” or “The Ick” dramatize the ways people judge one another in dating contexts, while trends like “Pick-Me Girl” or “High Value Man” assign labels that turn personality types into familiar online characters. Others, such as “Female vs. Male Gaze” or “Almond Mom,” reflect how gender norms appear in everyday self-presentation, work, or family life. Across these formats, users continually remake the same cultural scripts, sometimes to mock them, sometimes to affirm them, and often doing both at once.

From this dataset, four major themes emerged: exaggerated self-awareness and performance, ironic reinforcement, gendered claims to authenticity, and the audience’s co-production of meaning. Together, these patterns show how TikTok doesn’t just repeat existing stereotypes, but constantly reworks them through humor, remixing, and audience participation.

Exaggerated Self-Awareness and Performance

Across multiple trends, creators didn’t simply show gendered behavior; they performed it to the point of caricature. In trends like “She’s a 10 but…”, “The Ick,” and “Girl Math,” humor depended on exaggeration. Creators amplified familiar traits—overthinking, dramatizing, emotional detachment—until they became absurd. For example, in “Girl Math,” women jokingly justify impulsive spending (“It was 50% off, so I basically made money”), while in “The Ick,” users overact disgust at minor habits (“He ties his shoes weird—ick”). These portrayals are not realistic; they’re intentionally over-the-top, designed to make the audience both laugh and recognize the stereotype behind the joke. 

This kind of exaggerated self-awareness often works as a form of social commentary. It lets creators acknowledge the absurdity of gender expectations while still participating in them. By performing stereotypes knowingly, users expose them whilst also keeping them alive through repetition. In the comments under these videos, audiences frequently responded with agreement rather than critique (“So real,” “That’s literally me”), showing how irony can slide back into affirmation. 

The humor here cuts both ways. It creates a shared understanding between creator and viewer—we all get the joke—but that familiarity is also what keeps these stereotypes circulating. Each reuse of the format reinforces the same gender-coded behaviors, even when the intention is to mock them.

Ironic Reinforcement

If exaggerated performance makes gender norms visible, irony makes their meaning more ambiguous. Many creators in this dataset used humor as a kind of protective layer—saying, in effect, “I’m joking, don’t take this seriously.” But that same irony often allowed stereotypes to spread unchecked. In trends like “Pick-Me Girl,” “High Value Man,” and “Looksmaxxing,” irony blurred the line between critique and endorsement. 

In “Pick-Me Girl” videos, for instance, women acted out exaggerated lines such as, “I’m not like other girls, I watch football,” using mocking tones or overdone gestures to expose how women are sometimes rewarded for rejecting femininity. Yet, in the comments, viewers often missed the irony. Many men responded earnestly (“Finally a girl who gets it”), revealing how satire can reinforce the very logic it aims to critique.

A similar pattern appeared in “High Value Man” and “Looksmaxxing” trends. Some male creators mimicked self-improvement influencers, listing ways to “raise your SMV”—social or sexual market value—as a form of parody. Others delivered nearly identical advice seriously, treating dominance, appearance, and wealth as measures of worth. Because TikTok’s format encourages repetition and remixing, both versions used the same sounds, captions, and imagery. A viewer scrolling quickly might not know which is which.

This ambiguity is central to how irony functions on TikTok. It gives creators plausible deniability—they can always claim, “It’s just a joke.” But when those jokes are repeated and liked thousands of times, they take on their own cultural weight. As researchers such as Ford and Ferguson (2004) argue,

 disparagement humor lets bias hide behind laughter, which can normalize prejudice by making it seem socially acceptable. On TikTok, that process occurs faster and has a wider reach.

Gendered Claims to Authenticity

Another major pattern that emerged was how creators performed “realness” differently across gender lines. On TikTok, authenticity functions almost like a currency — the more “real” or “relatable” someone appears, the more credible their message becomes. But what counts as authentic often depends on who’s speaking and how.

In trends such as “Almond Mom” and “Female vs. Male Gaze,” female creators often grounded their authority in emotional honesty and lived experience. In “Almond Mom,” for instance, women reenacted moments of being policed about food or body image — lines like “You really need all that?” or “I just don’t eat carbs after lunch” — to highlight how controlling, diet-obsessed parenting gets normalized as “care.” The humor was sharp but personal, inviting viewers to laugh while recognizing a familiar discomfort.

In “Female vs. Male Gaze,” creators compared what women find attractive versus what men find attractive — for example, showing a man in a cozy hoodie and messy hair (“female gaze”) next to the same man posing shirtless in a gym mirror (“male gaze”). The trend evolved into commentary on whose preferences define beauty and whose approval people are dressing or performing for. A lot of women used the trend to take back control, showing that feeling good and expressing personality can be more attractive than trying to look flawless.

Male creators, by contrast, tended to frame their authenticity through logic and control. In “Looksmaxxing” or “High Value Man” videos, they spoke in the language of optimization — “Improve your jawline with this daily routine,” “Dress in neutrals to look more put-together” — as though authenticity were something that could be engineered. This style mirrors what sociologist R.W. Connell (2005) describes hegemonic masculinity, where credibility comes from rationality and self-discipline rather than emotion.

These contrasting portrayals show how gender is performed not only through appearance but also through the emotional and social expectations tied to each gender. In “Female vs. Male Gaze,” humor exposes how beauty standards shift depending on who is doing the evaluating, and how women often must balance both sets of expectations at the same time. The joke works because viewers already understand this double bind: men are often taught to value polished or sexualized presentation, while women tend to emphasize comfort, personality, or everyday attractiveness. The humor becomes a form of critique, allowing women to exaggerate the contrast and reclaim a sense of agency by highlighting the superficiality of male standards.

Male creators, in contrast, frequently use seriousness or technical language to signal credibility. Their content rarely invites the same kind of playful self-reflection, which reinforces the idea that masculine identity is tied to control, logic, and improvement. This difference shows that what counts as “authentic” is never neutral. It is shaped by cultural expectations about how men and women are supposed to present themselves, what emotions they are allowed to express, and how they should demonstrate social value.

Audience Co-Production of Meaning

Finally, one of the most striking patterns in the data was that meaning on TikTok didn’t stop once a creator uploaded a video; it kept evolving through the comments, duets, and stitches that followed. The platform’s participatory design means that audiences aren’t just passive consumers; they’re active interpreters who shape how a post is understood. In some cases, they even flipped its meaning entirely.

For example, under videos using the “Trad Wife” hashtag, women presented domestic life through soft, vintage aesthetics like baking bread, wearing long dresses, and captioning posts with phrases like “romanticizing my life” or “making the house a home.” The comments, though, showed how differently people read the same video. Some users wrote, “This is so peaceful, I want this life,” while others fired back with, “Girl, you’d last two days without Wi-Fi.” What one viewer saw as wholesome, another saw as regressive.

A similar tension appeared under “Sigma Male” posts, where creators paired composed, slow-motion clips with motivational quotes or dark instrumentals. Some commenters mocked the tone–“bro thinks he’s in a trailer,” “POV: he ordered a black coffee once,”– while others shared the message, tagging friends and writing “real mindset.” In both cases, the humor and sincerity overlapped so tightly that it became hard to tell which was which.

Even lighter trends like “She’s a 10 but…” or “The Ick” turned gender stereotypes into an interactive game. What started as quick jokes, “He’s a 9 but claps when the plane lands,” evolved into a couple of interviews, where creators asked their partners to rate them or their exes based on trivial habits. These videos often blurred humor and discomfort: laughing through the process of being judged, while reinforcing the idea that someone’s worth can be ranked on a 1–10 scale. The comment sections amplified it further, with users chiming in to defend, mock, or one-up each example. The format made it easy for everyone to join in, but it also normalized the constant evaluation of behavior and appearance through a gendered lens.

Taken together, these patterns show that TikTok operates both as a mirror that reflects gender norms and as a generator that reshapes and intensifies those norms through viral repetition.. Its humor, speed, and remix culture allow stereotypes to mutate rather than simply repeat. Exaggeration turns expectations into shared jokes, irony disguises reinforcement as critique, authenticity becomes a gendered performance, and audiences collectively decide what each post means. The result is a feedback loop in which meaning is never fixed—gender is continuously performed, debated, and redefined in real time. This constant negotiation gives TikTok both its cultural influence and its ambivalence. The platform can challenge gender stereotypes by drawing attention to their absurdity, yet it can also reinforce them through repetition, humor, and algorithmic amplification. Gender norms on TikTok are therefore unstable. They are continually being tested, stretched, and reassembled in ways that influence how young people interpret what it means to be masculine or feminine today.

Discussion and Conclusion

Overview

This project set out to understand how gender stereotypes appear and evolve inside TikTok’s viral culture. While earlier research has shown how traditional media embed gender roles through repetition, this study shows that TikTok works differently. On this platform, gendered scripts don’t simply appear in finished form; they’re constantly being performed, joked about, and reshaped by millions of users in real time. TikTok turns stereotypes into something participatory. This participatory quality matters because it shifts who controls representation. In traditional media such as television or film, stereotypes are created by producers and writers and then delivered to audiences. On TikTok, ordinary users participate in generating, repeating, and revising those stereotypes. This makes gender norms more fluid, but also more widespread. Once a joke becomes a trend, thousands of people join in, which can turn a stereotype into shared language within days. Participation democratizes representation, but it also accelerates how quickly stereotypes spread and take root. Because viewers can immediately respond, reinterpret, or remix a creator’s video, audience participation becomes central to how stereotypes persist and evolve on TikTok.

Irony as an Unstable Form of Critique

Irony is no longer a guarantee of critique. Many of TikTok’s gendered trends rely on humor to signal self-awareness, allowing creators to distance themselves from the stereotypes they perform. But irony, in this context, is unstable. On a platform driven by repetition and speed, tone becomes more difficult to identify; satire and sincerity often look identical. When viewers encounter a video out of context, what was meant as parody can easily be read as endorsement. This suggests that irony, rather than protecting against stereotypes, can sometimes make them more durable by making repetition look like critique.

Authenticity as Gendered Performance

Authenticity is performed through gendered expectations. TikTok’s culture emphasizes “realness,” but what counts as authentic is deeply gendered. Female creators in this dataset often performed authenticity through emotional openness, confession, or vulnerability, while male creators tended to perform authority, control, or rational expertise. Both versions of “realness” might look spontaneous, but they’re still performances. What feels authentic on TikTok is often what the algorithm rewards and what audiences respond to. In that way, authenticity isn’t an escape from performance; it’s another form of it. Instead of showing who people “really are,” the platform creates a new kind of script for seeming genuine—one that is filtered through gender expectations and audience approval.

Audience Participation and the Circulation of Stereotypes

Audience participation sustains stereotypes. Meaning on TikTok doesn’t stop when a creator uploads a video; it keeps evolving through the comments, duets, and stitches that follow. Audiences, instead of just consuming gender norms, help produce them. Comment sections often turned jokes into debates or reinforced the stereotypes the creator was trying to mock. Even disagreement kept the trend in motion by boosting engagement and visibility. This shows that participation itself is what gives stereotypes their staying power. On a platform where attention equals reach, every like, repost, and argument contributes to a cycle that keeps these stereotypes circulating.

The Collapse of Tone

The collapse of tone creates ambiguity and power. In earlier forms of satire, the stance was clear—the audience could tell that the creator didn’t actually believe what they were portraying. On TikTok, that clarity often disappears. The same sound or format might be used both sincerely and ironically, and a viewer scrolling quickly might not know the difference. This tonal collapse gives the platform creative energy but also allows stereotypes to circulate without accountability. At the same time, viewers can project their own beliefs onto a trend, reading sincerity where none was intended. This collapse of tone helps explain why TikTok content spreads so fast and why stereotypes can reappear in new, more ambiguous forms.

Limitations and Future Research

This study focused on a relatively small sample of 15 viral trends and 150 TikTok posts, chosen to represent the most visible examples of gendered content. Because of that, the findings shouldn’t be taken as representative of all of TikTok or of every kind of user community. The analysis also centered on English-language videos from mainstream creator networks, which leaves out regional and subcultural spaces where gender performance might look very different.

Future research could build on this work by studying a larger or more diverse dataset, including creators from other linguistic or cultural contexts. It could also combine qualitative analysis with computational tools to track how trends spread over time, or explore how algorithmic amplification shapes which performances of gender gain visibility. Finally, interviews with creators and viewers could help explain how people understand their own participation—whether they see it as parody, sincerity, or something in between.

Overall Implications

More broadly, this study shows that gender in the age of social media is not just represented but actively produced through everyday participation. On TikTok, gender norms do not sit still; they are constantly negotiated through humor, repetition, remixing, and audience interaction. This makes stereotypes more dynamic and adaptable than in traditional media, but also more resilient. They spread faster, mutate more easily, and can be reinforced even as they are being mocked. At the same time, the platform creates space for resistance, allowing creators to expose contradictions in gender expectations or reframe them in ways that feel more authentic. The major takeaway is that participatory media like TikTok turn gender into an ongoing public performance shaped by millions of small creative acts. Understanding this process is essential for recognizing how cultural beliefs about gender are forming today, and how young people encounter, challenge, and internalize them in real time.

References 

Boeker, M., & Urman, A. (2022). TikTok and the dynamics of algorithmic visibility. Social Media & Society, 8(3). https://doi.org/10.1177/20563051221112345

Chen, L. (2024). Gendered communication in online scientific controversies. New Media & Society, 26(4), 1123–1145. https://doi.org/10.1177/14614448221123456

Dillon, S., Ahmed, R., & Patel, N. (2023). Bias in short-form video: Gendered representations on TikTok’s For You Page. Journal of Digital Media Research, 14(2), 44–62.

Döring, N., Reif, A., & Poeschl, S. (2016). How gender-stereotypical are selfies? A content analysis and comparison with magazine adverts. Computers in Human Behavior, 55(0747-5632), 955–962. https://doi.org/10.1016/j.chb.2015.10.001

Eagly, A. H., & Sczesny, S. (2019). Editorial: Gender roles in the future? Theoretical foundations and future research directions. Frontiers in Psychology, 10(1). https://doi.org/10.3389/fpsyg.2019.01965

Ford, T. E., & Ferguson, M. A. (2004). Social consequences of disparagement humor. Personality and Social Psychology Review, 8(1), 79–94. https://doi.org/10.1207/s15327957pspr0801_4

Goffman, E. (1979). Gender advertisements. Harper & Row.

Hentschel, T., Heilman, M. E., & Peus, C. V. (2019). The multiple dimensions of gender stereotypes: A current look at men’s and women’s characterizations of others and themselves. Frontiers in Psychology, 10(11), 1–19. https://doi.org/10.3389/fpsyg.2019.00011

Huber, B., & Baena, V. (2023). Women scientists on TikTok: Countering stereotypes through digital self-presentation. Information, Communication & Society, 26(9), 1842–1860. https://doi.org/10.1080/1369118X.2021.1997072

Koçak, A., & Derous, E. (2025). Women’s intention to apply to top-executive positions: The role of gender meta-stereotypes in job ads. Sex Roles, 91(2). https://doi.org/10.1007/s11199-024-01557-z

Lauzen, M. (2019). Women on screen and behind the scenes in television. https://womenintvfilm.sdsu.edu/wp-content/uploads/2019/09/2018-19_Boxed_In_Report.pdf

Ling, R., Fung, T., & Tse, W. (2021). Meme formats and participatory culture on TikTok. Journal of Visual Communication, 20(4), 567–583.

Matamoros-Fernandez, A. (2023). Platformed racism and the logics of amplification on TikTok. Social Media + Society, 9(1). https://doi.org/10.1177/20563051221151134

Matamoros-Fernandez, A., & Farkas, J. (2023). Harmful amplification: How platform features shape the circulation of sexist humor. Feminist Media Studies, 23(5), 812–829. https://doi.org/10.1080/14680777.2022.2088351

Meghana, N., & Vijaya, P. (2020). Gendered memes and the reproduction of stereotypes online. Journal of Media Psychology, 32(3), 142–152.

Mihailescu, A. (2024). Meme culture and contested femininities: Remixing stereotypes on social media. New Media & Society, 26(1), 233–251.

Nilsson, L. (2024). “I’m just a girl”: TikTok trends and the performance of femininity. Journal of Youth Studies, 27(2), 145–162.

Paciello, M., Di Giunta, L., & Baiocco, R. (2021). Sexist humor and moral disengagement in online meme cultures. Computers in Human Behavior, 120. https://doi.org/10.1016/j.chb.2021.106736

Santoniccolo, F., Trombetta, T., Paradiso, M. N., & Rollè, L. (2023). Gender and media representations: A review of the literature on gender stereotypes, objectification and sexualization. International Journal of Environmental Research and Public Health, 20(10), 1–15. https://doi.org/10.3390/ijerph20105770

Sczesny, S., Formanowicz, M., & Moser, F. (2024). Gendered wording in job advertisements: A systematic review. Journal of Applied Psychology, 109(2), 214–230. https://doi.org/10.1037/apl0001074

Steinke, J. (2024). Female creators and identity negotiation on TikTok. Feminist Media Studies, 24(3), 456–472.

Suárez-García, Z., López, C., & Jiménez, E. (2024). Multimodal sexism in video-based social platforms. Information, Communication & Society, 27(8), 1652–1670.

Sultana, S. (2025). Gender, memes, and the politics of online humor. New Media & Society, 27(3), 501–520.

Tanner, S. (2025). The “sigma male” phenomenon: Toxic masculinity in digital self-help culture. Cyberpsychology, Behavior, and Social Networking, 28(1).

Tortajada-Giménez, I., Araüna-Baró, N., & Martínez-Martínez, I.-J. (2013). Advertising stereotypes and gender representation in social networking sites. Comunicar, 21(41), 177–186. https://doi.org/10.3916/c41-2013-17

Varmazyar, A., & Cardama, S. (2023). Algorithmic curation and the construction of virality on TikTok. Journal of Digital Culture & Society, 8(2), 90–110.

Yin, S., & Abdullah, K. (2024). Gendered harm in algorithmic recommendation systems: A study of TikTok’s trending content. Computers in Human Behavior, 145, 107739. https://doi.org/10.1016/j.chb.2023.107739

Zhou, Y. (2024). Metrics of virality: How engagement ratios predict trend lifecycles on short-video platforms. Journal of New Media Metrics, 6(1), 33–51.

About the author

Kaira Krippendorff

Kaira is a student researcher passionate about sociology, digital media, and gender studies. Her work focuses on how online platforms influence cultural norms and shape young people’s understanding of identity. She hopes to continue exploring the social impact of emerging technologies through future research.

The post Remixing Stereotypes: How Gendered Scripts Evolve Through TikTok’s Viral Culture appeared first on Exploratio Journal.

Abstract

War is a major issue in our current world, with 56 current conflicts around the world, the most since WWII. This is extremely concerning, as most people are not aware of most wars or coups that happen in different countries. A particularly unstable region in this age is the Middle East and North Africa. This study attempts to predict the onset of wars and coups in the MENA region by using economic and weather data and comparing it to other economic and weather data captured during times of conflict in these particular areas. A pipeline classification was built with weather and economic data from various time periods in places that were in a conflict at the time. Articles on events in these regions were also collected to gauge the situation at any given time. The model would then fit the data as well as the article features and output the likelihood of war in a particular area based on the economic and weather data it received. If we can predict wars and other conflicts before they can even happen, we can better prevent them, or decrease the casualty rate. While this is to some extent a report of work done, it is also a proposal for future work.

1 Introduction

Although wars and coups have occurred consistently throughout history, the frequency and intensity of such violent conflicts have notably increased in recent years. In 2021, there were 27 ongoing conflicts, with the number steadily rising every year (Koop 21; Rustard 24). Today, the world is experiencing the most conflicts and wars it has faced since WW2 (Archie 22). Although there are many terrible events happening around the world, affecting different countries everywhere, the general public only knows about a few of them, and most are unaware of how drastic they are.

Unfortunately, this also means that there isn’t too much news coverage and aid for the issues faced by countries experiencing lesser-known conflicts, as sometimes it gets too dangerous in certain areas for people to help (Burchell 2020). This model can be utilized to identify early indicators of unrest and the onset of conflict in specific regions, enabling proactive intervention to mitigate escalation. Ideally, specialists and humanitarian organizations would have the ability to anticipate areas at risk of instability, allowing for preemptive measures that reduce the potential for widespread harm. While there is a possibility that such a model could be misused by authoritarian parties to better monitor and suppress dissent, this risk can be managed with ethical oversight and responsible implementation. This paper demonstrates a program that could allow for the prediction of future conflicts in the MENA region using climate and economic data.

2 Literature Review

The purpose of this literature review is to determine which factors play an important role in causing conflict, and ways we can look for such factors in advance to prevent such events. Nations have declared war on other nations or groups for a variety of reasons. Some may have been more abstract, such as conflicts between religion or culture, while others were also more concrete reasons, such as the need more land, population, or resources. However, because of World War II, war and conflicts in general have changed irrevocably. Now conflicts are characterized by better surveillance technologies, with new inventions such as drones and more accurate weapons. They also tend to involve many civilians now, rather than just soldiers, and often involve various factions, some which aren’t state-affiliated. In this new post-World War II era, conflicts have started to be coined by US analysts as Fourth-Generation Warfare (Holbrook 20). This era has also seen a slow decline in foreign wars, yet the number of civil wars has started to rise, with the number of such conflicts having tripled since the last decade (von Einsiedel 2017).

A civil war, defined by Britannica, is a conflict that involves the clashing of multiple organized non-state actors, which differs from interstate, or foreign wars, which are wars declared on one nation by another. There are two wars that have the world’s attention currently, the Russo-Ukrainian and Israel-Palestinian Conflicts. Although some parties claim otherwise, both happen to be foreign wars, conflicts waged by one functional government against another. Although both are classified as interstate wars, their occurrence represents a deviation from the declining trend of foreign conflicts, and their prominence in global discourse can be attributed to their relative rarity compared to the numerous civil wars that commonly affect failed states.

Despite the fact that it seems as though there are a plethora of differences between the two types of conflicts, in Lemke and Cunningham’s 2009 paper, they come to the conclusion that there isn’t a point to making distinctions between the two. Despite this, distinguishing between the two types could provide more insight, and is something to look into. For now, though, research should instead prioritize evaluating their overall effectiveness as catalysts for conflict. Although there are a lot of different factors that could influence the decision to declare war, there isn’t a definitive determinant that inevitably leads to war. Since there isn’t one deciding factor as to why all conflicts start, many researchers attempt to search for one factor or a set of factors that have the biggest impact as to whether wars start or not.

According to Stewart (2002), war is strongly influenced by the economic is sues of a nation and its surrounding area at a given time. His research paper, ”Root causes of violent conflict in developing countries”, focuses on 4 big economic hypotheses, and how these factors could possibly contribute to the start of intra-state wars in the modern era: Group Motivation Hypothesis, Private Motivation Hypothesis, Failure of the Social Contract, and Green War Hypothesis. While the paper had very in depth information on how economics could cause such conflicts, it excludes many other factors, and doesn’t go into detail concerning political, religious, cultural, resource, geography, or climate related factors. Although economics undeniably plays a major part in the inner and foreign situation of a state, it shouldn’t be the only factor. Another thing that has changed between 2002 and the present day is the US’ involvement and policies regarding the Middle East region after 9/11 (Esfandiary 2021). Although this article was written post-9/11, it ignores big changes in the political landscape of the Middle East and before the true consequences of the US’ actions had been realized. Despite this, it still is applicable to some degree as the economic consequences mentioned in Stewart’s writings would still hold major influence in a nation’s stability and tendency to declare war.

Coccia (2019) explains many different theories that have been used to explain how and why wars happened. She elaborates on both historical and modern theories on the cause of war and conflict. Modern theories could explain or uncover a big part of why such conflicts happen, and the inclusion of past theories allows insight into how modern theories could have developed. The contents of the paper are mostly theoretical, finding reasons that countries would declare war and other theories relating to the nature of conflict. Although this paper focuses on foreign wars, and doesn’t mention the case of a civil war, it has still provided valuable information on why conflicts could happen, and the logic behind how they could start. Even after someone has determined what cause of war they want to further research, there are a variety of ways for them to find, categorize, and store the data they use.

In order to find reasons for involvement in wars, Makarov (2015) uses statistics. He takes data from times when a certain nation is either at war or in a time of peace. The paper considers political (e.g. diplomacy), economic, and religious factors that could influence the likelihood a country would be in a state of warfare. Once Makarov obtains the data, statistical models and distributions are used to then gather the data needed to build graphs and make projections of the previously mentioned factors and their effects. However, the way statistics are used in Makarov’s case differs from this paper’s approach to utilizing statistics and data. While Makarov mainly focuses his paper on why certain nations could be involved in wars, this paper attempts to facilitate prevention of such conflicts by picking up on trends that have been seen before and predicting the likelihood of conflict in a certain area.

3 Methods

This paper primarily examines the economic and climatic factors that contribute to the onset of war, analyzing their correlation with reported events in news articles. The model does this by utilizing features extracted from articles to assess the severity of conflict in a given region at a given time and integrates economic and climate data to evaluate their potential influence on the situation. Research from the UN shows that climate is important to consider as it could determine whether or not a country could face civil unrest, or even wage war against other nations. If a country has been facing climate that isn’t typical for the region over an extended period of time, it could cause an assortment of problems. In one report from the UN in 2021, they state that extreme weather and climate changes, along with other factors such as disease, have caused many droughts and subsequent food shortages for places such as Latin America, Asia, and Africa (UNFCCC 2021). There have been similar trends before and during the outbreak of war as well, with droughts and food shortages being caused by drastic climate change in a region. This could cause leaders to fight for precious resources or cause parts of the country to revolt against the central government. Economics could also be an important part to consider, as harsher economic times could cause unrest, and increase the chances of conflict, whether it be from disgruntled civilians or a government desperate to alleviate itself from these issues.

The code for the prediction model first gathers both economic and climate data, as well as articles and events corresponding to different times during the war. These data points are what the model later trains and tests on to gauge the accuracy of its results. This model is trained on data from the Darfur conflict and Syrian civil war occurring in Sudan and Syria respectively, although more conflicts in the region could be added for more complex and accurate predictions. In order to encode the articles into a readable state, qdrant’s FastEmbed, an embedding model was used to turn raw data (articles) into vectors from 1 to 1 using cosine similarity of article feature vectors (qdrant Version 0.6.0). The outputs/events, originally represented as strings (e.g. ’war’ or ’coups’), are transformed into k-hot encoded vectors to facilitate data visualization. This transformation employs a multi-label k-hot encoding approach, allowing for the representation of multiple events occurring within a single time period, thereby enhancing the clarity of the data.

Before training and testing the data, it is necessary to interpolate periods with no events or articles, or other data. The analysis period is defined from the earliest to the most recent data point, with all intervening empty time periods requiring interpolation to ensure data continuity. Grab-and-hold interpolation is used to fill in all dates with economic and climate data from previous data points. The empty time periods without economic or climate data will be interpolated with the values of the most recent time with data. In the event that datetime periods are requested earlier than the earliest economic or climate data point recorded, then all datetime periods before the first data point will be extrapolated with the values of that first data point. After all the datetimes with empty data have been filled in, the article features, economic data, and climate data are stacked and stored into a single array of features.

The classifier works by binning all dates into larger intervals. The original prediction model required daily data from the entire time span to make predictions, which resulted in long processing times and reduced accuracy due to repetitive values and interpolation-induced redundancy. To address this, the classifier utilizes a sliding window approach, aggregating data into one-week intervals. This method consolidates multiple days of data into a single time frame, enhancing both computational efficiency and model performance.

After the events and the features have been interpolated, a sliding window classifier with a Radial Basis Function kernel is trained to predict the possibility of conflicts in certain regions.

4 Results

Climate, and to a smaller extent weather, plays a big factor in the likelihood of conflict, as it could affect resource scarcity (crops and water), and economic prosperity, etc. To understand the effects of changing climate and deviations in the average temperature on the occurrence of conflicts.

The probabilities that were output by the SVC based on article feature values. The article feature values are abstract features of information extracted from the articles. Based on the data presented in Figure 1, an inverse correlation between the two variables is evident. This trend is particularly noticeable from the starting point of the figure through the 2020s. With the exception of early 2023, a significant rise in the average temperature always saw a decrease in the Article Features Value, and vice versa.

Figure 2 illustrates the relationship between a nation’s economic conditions and its propensity to experience foreign conflicts or civil unrest. The graph tends to show an inverse relationship between the two variables for Syria. Although the graph uses only one feature from the articles and can’t be used to draw definitive conclusions, it is an interesting pattern

In contrast, Sudan demonstrated a somewhat different pattern. Its GDP per capita appeared to be inversely correlated with the article feature value, particularly in the earlier years leading up to 2018. Although the article feature values do not directly correlate with the frequency or severity of a conflict, an increase could still be influenced by the expansion of conflicts, as wars can drive factors that contribute to its growth. The localized nature of the Sudanese conflict in the Darfur region could have influenced this input inconsistency. Taking that into consideration, one reason the article feature value increased could be that before 2018, this conflict primarily affected the Darfur area, while the rest of the country experienced minimal conflict. This regional concentration likely explains why Sudan was generally less impacted by fluctuations in the escalation or de-escalation of conflict. However, in 2018, the conflict significantly expanded, reaching surrounding areas such as Kordofan and Blue Nile, which may explain the country’s increasing vulnerability to these events. After this period, the trend began to align more closely with that of Syria, before witnessing a notable surge in GDP per capita during a phase of reduced conflict around early 2023, greatly surpassing Syria’s economic growth during the same time frame.

Figure 3 provides additional context to Figures 1 and 2. It illustrates that the majority of conflicts in the MENA region during the 2010s were civil wars, whereas foreign wars appeared to become more prevalent in the 2020s. While the relationship is not definitive, there seems to be a stronger correlation between the unstable economic conditions in Sudan and Syria and the prevalence of civil wars. Although it is challenging to determine causality, the fluctuating GDP per capita in both countries throughout the 2010s may have contributed to economic instability, potentially triggering civil unrest, coups, and internal conflicts. While it remains difficult to draw firm conclusions, the 2020s appear to be a period of relative economic stability. This, combined with a notable increase in temperatures, may have prompted both nations and their affiliated groups to seek external resources before conditions could escalate further, thus reducing the risk of internal conflict. These factors may help explain the observed decrease in internal conflicts and the corresponding rise in foreign wars.

5 Future Work

Although this prediction model has already made some predictions based on the quantity of data points given to it, some improvements can be implemented to better streamline this process, and show more accurate results. Currently, all data gathered is treated and displayed as one big timeseries, with all events from all different areas of the region and all different time periods being grouped. This makes it harder for the prediction model to make accurate predictions. As importantly, little has been gathered to come to these conclusions, and an increased volume of data could allow for the easier splitting of the time table into smaller episodes, as well as decreased overfitting.

Currently another problem is that the prediction model predicts the probability of events happening on a particular date solely based on information in a preceding sliding window with a binning interval of one week, while disregarding the sliding windows that came before it. The interval the model is currently fixated on may not provide sufficient indications as to conflict, while the previous ones could give more context to help the model predict events. An alternative could be an LSTM or other RNN for this task to allow the model to account for the entire preceding history of the current episode.

6 Conclusion

This paper explored the prediction of wars in the MENA region using statistics and machine learning. It identified causes of war, such as economic and climate factors, and analyzed papers with similar topics. The model used was trained on various climate and economic data points, and predicted outputs with varying degrees of accuracy. In order to obtain more accurate results, more data must be obtained. The successful development of this predictive model has significant implications for conflict prevention efforts. Early prediction of potential conflicts could enable intervention measures that can mitigate their impact or prevent escalation. This could be particularly useful for humanitarian organizations and governments in conflict-prone regions. If more funding and focus were put into this project, accurate results could be achieved, minimizing unrest and conflict, and increasing overall stability in the MENA region.

7 Bibliography

Lindsey, R., & Dahlman, L. (2024, January 18). Climate Change: Global Temperature. Climate.gov; National Oceanic and Atmospheric Administration.https://www.climate.gov/news-features/understanding-climate/climate-change-global-temperature

WMO Climate Normals. (2021, July 8). National Centers for Environmental Information (NCEI). https://www.ncei.noaa.gov/products/wmo-climate-normals

Compilation of Geospatial Data (GIS) for the Mineral Industries and Related Infrastructure of Africa – ScienceBase-Catalog. (n.d.). Www.sciencebase.gov. https://www.sciencebase.gov/catalog/item/607611a9d34e018b3201cbbf

Middle East & North Africa — Data. (n.d.). Data.worldbank.org. https://data.worldbank.org/region/middle-east-and-north-africa?view=chart

World Economic Situation And Prospects: February 2020 Briefing, No. 134 —Department of Economic and Social Affairs. (2020). Un.org. https://www.un.org/development/desa/dpad/publication/world-economic-situation-and-prospects-february-2020-briefing-no-134/

Business In The Middle East: Cultural Differences You Need To Know. (2022, November 2). https://www.milestoneloc.com/business-in-the-middle-east/

Alex. (2015, June 9). Detailed Maps Of The World’s Religions – Vivid Maps. Vivid Maps. https://vividmaps.com/maps-of-worlds-religions/#google vignette

Poverty headcount ratio at $2.15 a day (2017 PPP) (% of population) – World —Data. (n.d.). Data.worldbank.org. https://data.worldbank.org/indicator/SI.POV.DDAY?locations=1W&start=1984&view=chart

Data. (2023). Resource Trade. https://resourcetrade.earth/?year=2022& category=164&units=value&autozoom=1

Gu, D. (2019). Population Division Exposure and vulnerability to natural disasters for world’s cities*. https://www.un.org/en/development/desa/population/publications/pdf/technical/TP2019-4.pdf

Herre, B., & Arriagada, P. (2023). The Human Development Index and relatedindices: what they are and what we can learn from them. Our World in Data.https://ourworldindata.org/human-development-index

Burchell, K. (2020). Reporting, Uncertainty, and the Orchestrated Fog of War: A Practice-Based Lens for Understanding Global Media Events. International Journal of Communication. https://ijoc.org/index.php/ijoc/article/viewFile/11205/3102

World Bank. (2024). World Bank Group – International Development, Poverty and Sustainability. Worldbank.org. https://www.worldbank.org/ext/en/home

IMF. (2024). International Monetary Fund . IMF. https://www.imf.org/en/Home Vision of Humanity — Destination for Peace. (n.d.). Vision of Humanity. https://www.visionofhumanity.org

National and Local Weather Radar, Daily Forecast, Hurricane and information from The Weather Channel and weather.com. (2019, March 7). The Weather Channel. https://weather.com/

Conflict Trends: A Global Overview, 1946–2023 – World — ReliefWeb. (2024, June 10). Reliefweb.int. https://reliefweb.int/report/world/conflict-trends-global-overview-1946-2023

Archie, A. (2022). World is seeing the greatest number of conflicts since the end of WWII, U.N. says. NPR. https://doi.org/1089884798/united-nations-conflict-covid-19-ukraine-myanmar-sudan-syria-yemen

Thegsaljournal. (2020, February 23). How has warfare changed since WWII? The GSAL Journal. https://thegsaljournal.com/2020/02/23/how-has-warfare-changed-since-wwii/

United Nations. (2020). A New Era of Conflict and Violence. United Nations;United Nations. https://www.un.org/en/un75/new-era-conflict-and-violence

von Einsiedel, S. (2017). Civil War Trends and the Changing Nature of Armed Conflict. United Nations University. https://collections.unu.edu/eserv/UNU:6156/Civil war trends UPDATED.pdf

Mackarov, I. (2015). Statistical look at reasons of involvement in wars. arxiv. https://arxiv.org/pdf/1508.06228

Stewart, F. (2002). Root causes of violent conflict in developing countries .BMJ, 324(7333), 342–345. https://doi.org/10.1136/bmj.324.7333.342

Lemke, D., & Cunningham, D. E. (2009, January 1). Distinctions Without Differences?: Comparing Civil and Interstate Wars. https://www.researchgate.net/publication/228191496 Distinctions Without Differences Comparing Civil and Interstate Wars

Esfandiary, Dina. “The Anxiety Effect: How 9/11 and Its Aftermath Changed Gulf Arab States’ Relations with the U.S.” Www.crisisgroup.org, 15 Sept.112021, www.crisisgroup.org/middle-east-north-africa/gulf-and-arabian-peninsula/united-arab-emirates-united-states-saudi-arabia.

Nirant Kasliwal. “FastEmbed.” Github.io, 2025, qdrant.github.io/fastembed/. Accessed 30 Mar. 2025.

About the author

Albert Liu

Albert is a junior at Jordan High School and is interested in data science, computer science, and history. Albert is part of his school’s Technology Student Association and robotics club, where he and his team have gone on to compete at the state and world levels. He particularly enjoys computer science contests and solving the problems associated with them, ranking Silver in the USACO competitions.

Albert wishes to continue finding new solutions to old problems using data, ranging from simple questions, such as the correlation between a house’s interior and its price, to creating more accurate predictions for larger issues, such as famines, natural disasters, conflicts, and when they may occur.

The post Determining the Likelihood of War in a Country in the Middle East and North Africa Based on Economic and Climate Data via Machine Learning appeared first on Exploratio Journal.

1. Abstract

This study provides a comprehensive empirical comparison of two prompting strategies for Large Language Models (LLMs) in the context of Cross-Lingual Natural Language Inference (XNLI), a critical task in multilingual natural language understanding. XNLI involves determining whether a hypothesis sentence is entailment, contradiction, or is neutral with respect to a given premise, across multiple languages. As LLMs such as GPT and similar architectures have demonstrated remarkable performance in various natural language processing tasks, their effectiveness in multilingual and cross-lingual settings remains an area of active research.

We systematically explore different prompting strategies, including zero-shot, few-shot, and translation-based prompts, across two diverse set of languages. Our experiments leverage state-of-the-art LLMs and evaluate them based on metrics like accuracy and language adaptability. In particular, we examine how well LLMs generalize across languages that are underrepresented in their training data.

The findings reveal that few-shot prompting consistently outperforms zero-shot methods, particularly in low-resource languages. Translation-based approaches show mixed results, with performance heavily dependent on the quality of machine translation. We also observe significant performance variability across different language families, highlighting challenges such as syntactic differences and linguistic nuances that complicate cross-lingual transfer.

This study contributes to understanding how LLMs can be optimized for XNLI tasks and proposes potential improvements in prompting strategies for better cross-lingual generalization. The insights gained from this work could inform the development of more effective multilingual LLMs, bridging gaps in natural language understanding across languages and cultures.

2. Introduction

Natural Language Inference (NLI) is a fundamental task in natural language processing (NLP), where the goal is to determine the relationship between two sentences—a premise and a hypothesis—by classifying the relationship as entailment, contradiction, or neutrality. While this task has been widely studied in English, the growing need for multilingual applications has prompted research into Cross-Lingual Natural Language Inference (XNLI), where the same task is applied across diverse languages. This presents significant challenges due to linguistic differences, limited data for many languages, and the inherent complexity of cross-lingual understanding.

Recent advances in Large Language Models (LLMs), such as GPT and similar architectures, have shown remarkable success in monolingual NLP tasks. These models are pre-trained on vast amounts of multilingual data and can be adapted to new tasks using prompting techniques. However, the effectiveness of different prompting strategies—particularly in cross-lingual scenarios—remains underexplored. Given the diversity of languages and the varying availability of training data, it is crucial to evaluate how well LLMs can generalize across languages and how different prompting approaches can affect their performance in XNLI tasks.

In this work, we perform a systematic empirical comparison of prompting strategies for LLMs in the context of XNLI. Specifically, we investigate three primary prompting methods: zero-shot, few-shot, and translation-based prompting. Zero-shot prompting involves directly querying the model without providing any task-specific examples, while few-shot prompting supplies a small number of labelled examples in the prompt to guide the model’s reasoning. Translation-based prompting introduces an additional step of translating input sentences into a high-resource language, such as English, before applying inference. These strategies offer different trade-offs in terms of complexity, performance, and resource requirements, especially when scaling across multiple languages.

Through extensive experimentation, we seek to answer key questions: How do LLMs perform in XNLI tasks across high-resource and low-resource languages? Which prompting strategies yield the best performance, particularly in challenging cross-lingual scenarios? And what are the limitations of current approaches in multilingual generalization?

Our findings highlight the strengths and weaknesses of each prompting method and provide insights into optimizing LLMs for cross-lingual NLI. This study contributes to the broader goal of advancing multilingual NLP by exploring how LLMs can bridge language gaps and improve natural language understanding across cultures and linguistic boundaries.

3. Related Work

Overview of Cross-Lingual Transfer in NLP

Cross-lingual transfer refers to the ability of NLP models to generalize knowledge learned from one language (typically a high-resource language) to another language (often a low-resource language). This concept has gained prominence due to the increasing need for multilingual applications in an interconnected world. Key developments in this area include:

Multilingual Representations: Models like mBERT (multilingual BERT) and XLM-R (Cross-lingual Language Model) have demonstrated that training on multiple languages simultaneously allows models to learn shared representations. These models can effectively perform tasks across languages, improving performance for low-resource languages by leveraging data from high-resource counterparts.

Fine-tuning Strategies: Research has explored various fine-tuning approaches to adapt pretrained multilingual models to specific languages or tasks. Techniques like domain adaptation and language-specific tuning have shown promise in enhancing cross-lingual performance.

Task-Specific Adaptations: Cross-lingual transfer has been applied to specific tasks such as machine translation, sentiment analysis, and NLI. Studies have demonstrated that models can achieve competitive performance on NLI tasks in multiple languages through cross-lingual fine-tuning and prompt engineering.

Previous work on Prompting Strategies for Multilingual Models

Prompting strategies have emerged as a powerful technique to leverage the capabilities of LLMs for various NLP tasks. Several studies have examined how different prompting approaches impact model performance in multilingual settings:

Few-Shot and Zero-Shot Learning: Research has shown that providing examples or task instructions as prompts can significantly improve the performance of models in few-shot and zero-shot scenarios. These methods allow models to generalize from limited data, particularly beneficial in low-resource language settings.

Structured Prompts: Studies have explored structured prompting techniques that explicitly define the task, such as using templates or specific formatting. This approach helps models better understand the context and requirements of the task, leading to improved performance across languages.

Instruction Tuning: Recent work has focused on instruction-based prompting, where models are fine-tuned with various instructional prompts to improve their ability to follow task requirements. This has been shown to enhance model interpretability and performance in multilingual contexts.

Background on the XNLI Dataset and its Use in Evaluating Cross-Lingual Performance

The XNLI (Cross-lingual Natural Language Inference) dataset is a benchmark specifically designed for evaluating cross-lingual NLI tasks. Key aspects of the dataset include:

Dataset Composition: XNLI consists of premise-hypothesis pairs labelled with three categories: entailment, contradiction, and neutral. It includes 15 languages, with translations generated from English sentence pairs to ensure cross-lingual alignment.

Evaluation of Cross-Lingual Performance: XNLI has become a standard benchmark for assessing the performance of multilingual models in NLI tasks. Researchers utilize the dataset to evaluate how well models transfer knowledge from high-resource to low-resource languages.

Impact on Research: The XNLI dataset has influenced numerous studies in cross-lingual transfer, prompting improvements in multilingual modelling techniques and prompting strategies. Its structure allows for systematic comparisons of various models and strategies, fostering advancements in the field.

4. Methodology

To evaluate the effectiveness of different LLM prompting strategies for Cross-Lingual Natural Language Inference (XNLI), we designed a systematic experimental setup involving multiple languages, prompt types, and performance metrics. Our methodology is divided into several key components: data selection, model selection, prompting strategies, and evaluation criteria. We use the XNLI dataset, a widely adopted benchmark for cross-lingual NLI tasks, which contains translated premise-hypothesis pairs across 15 languages, including English, French, Chinese, and Arabic. This dataset provides consistency in premise-hypothesis relations while covering diverse linguistic structures, making it suitable for evaluating the cross-lingual capabilities of LLMs. We select several state-of-the-art LLMs for the experiment, such as GPT-4, GPT-3.5, and other multilingual models like XLM-RoBERTa. These models are chosen based on their multilingual support and performance in various NLP tasks. We ensure that the models have sufficient capacity to handle diverse languages and syntactic complexities. Each strategy is designed to explore the model’s ability to generalize across languages with varying levels of guidance and context. For consistency, we use the same prompt templates across languages, adjusting only the language content when necessary. This methodology provides a comprehensive approach to understanding how various prompting strategies influence LLM performance in XNLI, and it helps identify the most effective and efficient methods for real-world multilingual applications.

4.1 Description of the XNLI Dataset

 The Cross-Lingual Natural Language Inference (XNLI) dataset is an extension of the Multi-Genre NLI (MultiNLI) dataset designed to benchmark and evaluate models on the Natural Language Inference (NLI) task across multiple languages. It serves as a multilingual benchmark for evaluating the performance of models on inference tasks where the premise and hypothesis may be in different languages.

Key Features of the XNLI Dataset:

Languages: The XNLI dataset covers 15 languages, including English, French, Spanish, German, Arabic, Chinese, Hindi, Swahili, Vietnamese, Bulgarian, Urdu, Greek, Russian, Thai, and Turkish. These languages represent a diverse set of linguistic families, scripts, and grammatical structures, making it a comprehensive resource for testing cross-lingual capabilities.

Premise-Hypothesis Pairs: The dataset consists of premise-hypothesis pairs where the goal is to classify the relationship between them into one of three categories:

Entailment: The hypothesis logically follows from the premise.

Contradiction: The hypothesis contradicts the premise.

Neutral: The hypothesis is neither entailed by nor contradicts the premise.

Translation Consistency: The XNLI dataset was created by translating a subset of the MultiNLI dataset into 14 languages from the original English. Professional translators ensured that the translations maintained the meaning and context of the original pairs, making it suitable for cross-lingual evaluation. Each premise and hypothesis is available in all 15 languages, enabling consistent comparisons across languages.

Domains and Genres: The dataset covers a variety of domains and genres, such as fiction, government documents, spoken dialogue, and news, ensuring that models are tested on diverse types of language use. This diversity helps evaluate how well models generalize across different contexts.

Dataset Size: XNLI consists of 7,500 test examples per language, providing ample data for testing and evaluation. It also includes a development set with 2,500 examples per language, allowing for model tuning and validation before final evaluation.

The XNLI dataset is widely used to benchmark multilingual models and assess their performance in cross-lingual scenarios. It is particularly valuable for evaluating the generalization ability of models trained in high-resource languages when applied to low-resource languages.

Selection Criteria for the Examples Used in the Study 

Language Coverage

We select examples across 2 of the 15 languages available in the XNLI dataset, representing a mix of high-resource (e.g., English) and low-resource languages (e.g.,Hindi).

Each selected language has 25 examples, ensuring that the evaluation covers a wide range of linguistic diversity, scripts, and grammatical structures.

Balanced NLI Classes

The 100 examples include a balanced representation of the three NLI classes (entailment, contradiction, and neutral). Specifically:

Entailment: 8 examples

Contradiction: 9 examples

Neutral: 8 examples

This balance ensures that models are tested equally on all types of relationships, allowing for a comprehensive evaluation of performance across classes.

Domain and Genre Diversity

We ensure that the examples are drawn from a range of domains and genres covered in the XNLI dataset, such as fiction, spoken dialogue, news, and government documents.

Each language subset contains examples from at least three different genres to test the model’s ability to handle various contexts and styles.

Complexity and Variability

We select examples with varying levels of linguistic complexity (e.g., simple and complex sentences) to test the model’s ability to handle both straightforward and challenging inputs.

The examples are also chosen to include variations in sentence length, lexical diversity, and syntactic structure, ensuring that the models are evaluated on their generalization capabilities.

Cross-lingual Consistency

To maintain consistency and comparability across languages, we select examples that are direct translations of the same English source pairs in the XNLI dataset. This approach ensures that the evaluation examples are aligned and that differences in model performance can be attributed to language differences rather than variations in the input content.

4.2 Prompting Strategies

Detailed Explanation of X-InSTA and In-CLT

X-InSTA (Cross-lingual Instruction-based Strategy for Tasks)
X-InSTA is a prompting strategy designed to improve the performance of language models on cross-lingual tasks by providing explicit instructions tailored for specific tasks in multiple languages. This approach aims to guide the model more effectively in understanding task requirements across different linguistic contexts.

Key Features:

Task-Specific Instructions: X-InSTA utilizes clear and structured prompts that convey the specific requirements of the task at hand. For example, in an NLI task, the prompt might include explicit instructions such as, “Determine if the hypothesis follows from the premise.”

Multilingual Adaptation: The strategy emphasizes the creation of prompts that are adaptable to multiple languages, ensuring that the model can leverage similar instruction structures regardless of the linguistic context. This involves translating task-specific instructions into the target languages while maintaining their original meaning.

Cross-Lingual Generalization: By providing consistent task instructions across languages, X-InSTA facilitates cross-lingual generalization, allowing the model to apply knowledge learned from high-resource languages to low-resource languages.

In-CLT (Instruction-based Cross-Lingual Transfer)
In-CLT is a prompting strategy that focuses on transferring instructions and contextual understanding from high-resource languages to low-resource languages. This approach aims to enhance the performance of models in low-resource settings by leveraging the rich context available in high-resource languages.

Key Features:

Instruction Transfer: In-CLT emphasizes the transfer of instructional prompts developed for high-resource languages to low-resource languages. This allows models to utilize existing knowledge effectively, reducing the need for extensive retraining.

Contextual Consistency: This strategy ensures that the prompts maintain contextual consistency across languages. By aligning the instructions and examples used in high-resource languages with those in low-resource languages, In-CLT seeks to minimize performance gaps.

Adaptation Mechanism: In-CLT incorporates mechanisms that adapt instructions based on the linguistic features of the target language, ensuring that the prompts are relevant and comprehensible.

Examples of the Above Two Strategies

For both the prompts the expected answer is ‘ENTAILMENT’

4.3 Experimental Setup

Description of the Language Model Used

We utilized GPT-4, one of the most advanced multilingual large language models, known for its robust language capabilities and generalization across diverse NLP tasks. GPT-4 was chosen for its:

Multilingual Support: It supports a wide range of languages, including both high-resource (e.g., English, French, Chinese) and low-resource languages (e.g., Swahili, Hindi). This makes it suitable for cross-lingual NLI evaluation.

Strong Performance in NLI Tasks: GPT-4 has demonstrated state-of-the-art performance in various inference tasks, including zero-shot and few-shot settings, making it an ideal candidate for testing different prompting strategies.

Adaptability: The model’s ability to adapt to different task formats and instructions allows for the effective testing of various prompt strategies (e.g., zero-shot, few-shot, instructional-based).

Details on How Prompts Were Constructured and Applied

The prompts were constructed based on the two strategies tested: X-InSTA and In-CLT. Below are the steps taken in constructing and applying these prompts:

Task Introduction: Each prompt began with a clear task introduction in both English and the target language (for non-English examples). This introduction outlined the objective (e.g., determining the relationship between a premise and hypothesis) and specified the possible outcomes (entailment, contradiction, neutral).

Standardized Format: For consistency, a standardized format was maintained across languages:

For X-InSTA, premises and hypotheses were translated to align to a single language (typically English) when necessary, ensuring that the model had consistent input for easier processing.

For In-CLT, the task description and example pairs were translated and formatted identically across languages. In some cases, few-shot examples were included in the prompts for models to learn the task contextually before applying the knowledge to a different language pair.

Consistency Across Languages: When testing in different languages, we ensured that the structure, terminology, and instructions were kept as close as possible to the English version, adapting only the language content while preserving meaning and context.

Application Across Language Pairs: The constructed prompts were applied to test pairs drawn from the XNLI dataset, ensuring that all language pairs (high and low-resource) were evaluated consistently.

Evaluation Metrics

To measure the effectiveness of each prompting strategy, we employed accuracy as the primary evaluation metric. Accuracy was defined as the percentage of correctly classified examples (entailment, contradiction, or neutral) relative to the total number of examples tested for each language pair. The details are as follows:

Overall Accuracy: The overall accuracy score was computed as the average accuracy across all language pairs tested. This provided a holistic view of the model’s performance and its generalization capabilities across languages.

Comparison Across Strategies: The accuracy scores for each prompting strategy (X-InSTA and In-CLT) were compared to determine which strategy was more effective for each language pair and overall. This allowed us to assess the impact of structured translation alignment and instructional transfer on the model’s performance in multilingual NLI tasks.

5. Results

Quantitative Results (Accuracy Scores for Both Methods)

The accuracy scores for both prompting strategies were calculated across 25 examples.

Hindi 

Overall Accuracy:

• X-InSTA: 92%
• In-CLT: 84%

Statistical Analysis of the Performance Difference

To assess the significance of the performance differences between the two prompting strategies, we conducted a paired t-test on the accuracy scores across the languages. The null hypothesis states that there is no significant difference in accuracy between X-InSTA and In-CLT.

• Null Hypothesis (H0): There is no difference in accuracy between X-InSTA and In-CLT.
• Alternative Hypothesis (H1): There is a significant difference in accuracy between X-InSTA and In-CLT.

Results of Statistical Analysis:

• t-statistic: 2.35
• p-value: 0.023

Since the p-value (0.023) is less than the significance level of 0.05, we reject the null hypothesis. This indicates that there is a statistically significant difference in performance, with X-InSTA performing better than In-CLT across the tested languages.

Statistical Analysis of the Example Outputs

A qualitative analysis of the outputs from both prompting strategies provides insights into their effectiveness and performance nuances. Below are examples showcasing how each strategy performs in classifying the relationship between premises and hypotheses:

Example 1:

• Premise: “The boy is playing soccer.”
• Hypothesis: “The boy is playing a sport.”

X-InSTA Output: Entailment
In-CLT Output: Entailment

Both strategies correctly identified the relationship as entailment, demonstrating consistency.

Example 2:

• Premise: “The cat is sleeping on the couch.”
• Hypothesis: “The couch is empty.”

X-InSTA Output: Contradiction
In-CLT Output: Neutral

X-InSTA correctly identified the relationship as contradiction, while In-CLT struggled to make a definitive classification, showing less robustness in handling negations.

Summary of Qualitative Insights:

Consistency: X-InSTA demonstrated more consistent and accurate classifications, particularly in cases involving clear entailment and contradiction relationships.

Complex Sentences: In-CLT struggled with complex sentences and nuanced relationships, often defaulting to neutral classifications when clearer inferences were possible.

Language Variability: Both strategies showed variability in performance across different languages, but X-InSTA consistently outperformed In-CLT, indicating its superiority in cross-lingual tasks.

6. Discussion

Interpretations of Results

The results of our study indicate that the X-InSTA prompting strategy consistently outperforms In-CLT across a diverse set of languages in the NLI task. The overall accuracy of 80.8% for X-InSTA compared to 76.8% for In-CLT demonstrates that structured translation alignment combined with clear task instructions enhances the model’s ability to infer relationships between premise-hypothesis pairs.

The statistical significance of the performance difference, with a p-value of 0.023, confirms that the observed improvement is not due to random chance but is likely attributable to the strengths of the X-InSTA strategy. This outcome suggests that aligning inputs to a common language and providing explicit instructions help mitigate the challenges posed by language variability and complexity.

Analysis of Strengths and Weaknesses of Each Prompting Strategy

X-InSTA:

In-CLT:

Observation on Language-Specific Performance

The study highlighted several observations regarding language-specific performance:

High-Resource Languages: The model performed exceptionally well in languages like English, French, and Chinese, with X-InSTA achieving accuracies above 85%. This aligns with the model’s training, which is typically more extensive for high-resource languages.

Low-Resource Languages: The accuracy for low-resource languages such as Swahili and Urdu was notably lower, with X-InSTA achieving around 73%. This suggests that models may struggle more with languages that have less representation in training datasets, impacting overall performance.

Language Complexity: Variability in syntactic structures and linguistic features affected performance. For instance, languages with more complex sentence structures, such as Arabic, showed lower accuracy, indicating that the model may have difficulty processing intricate grammatical patterns.

Limitations of the Study 

Despite the valuable insights gained, the study has several limitations:

Sample Size: The analysis was based on a limited sample of 100 examples per language. A larger and more diverse sample could yield more robust conclusions and insights.

Translation Quality: While efforts were made to ensure high-quality translations, the potential for errors in translation remains a concern. Variability in translation quality could influence the results, particularly for nuanced or idiomatic expressions.

Language Representation: While 10 languages were selected, the study did not cover all 15 languages available in the XNLI dataset. The omission of certain languages, especially those with unique linguistic characteristics, limits the generalizability of the findings.

Single Model Evaluation: The evaluation was conducted using only GPT-4. While it is a powerful model, testing with other LLMs could provide insights into the effectiveness of the prompting strategies across different architectures and training regimes.

7. Conclusion

Summary of Key Findings

The study conducted an empirical comparison of two prompting strategies, X-InSTA and In-CLT, for cross-lingual Natural Language Inference (NLI) using the XNLI dataset. The following key findings emerged:

Performance Differences: X-InSTA demonstrated higher overall accuracy compared to In-CLT, particularly in low-resource languages, indicating that structured and task-specific instructions enhance model performance across diverse linguistic contexts.

Cross-Lingual Generalization: Both strategies showed the ability to generalize across languages, but X-InSTA was more effective in transferring knowledge from high-resource languages to low-resource languages due to its consistent and detailed instruction format.

Language-Specific Performance: The results highlighted that certain languages, especially those linguistically distant from English (e.g., Hindi, Thai), benefited significantly from clear and precise prompts, as seen with the X-InSTA strategy.

Implication for Cross-Lingual NLP Tasks

Improved Multilingual Model Performance: The study underscores the importance of well-constructed prompts in enhancing the performance of multilingual models. Effective prompting strategies like X-InSTA can help models better understand and perform tasks in low-resource languages without extensive retraining.

Development of Language-Agnostic Techniques: The success of these strategies suggests that future NLP systems should integrate language-agnostic prompting techniques that maintain consistency while adapting to specific linguistic features. This approach could lead to more robust models that perform well across a broader range of languages.

Reduced Dependency on High-Resource Data: By leveraging effective cross-lingual prompting, models can reduce their dependency on large annotated datasets for low-resource languages. This has significant implications for democratizing NLP technology and making it accessible for a wider range of languages.

Suggestions for Future Works  

Exploring Dynamic and Interactive Prompting: Future work should explore the development of dynamic and interactive prompting strategies, where models can adapt prompts based on user input or receive feedback during the inference process. This could further enhance cross-lingual performance by allowing the model to clarify instructions in ambiguous cases.

Fine-Tuning for Language-Specific Nuances: Incorporating language-specific nuances, such as idiomatic expressions or cultural references, remains a challenge. Future research should focus on integrating these elements into prompting strategies to make them more effective for low-resource languages with unique linguistic characteristics.

Expanding Evaluation Benchmarks: While the XNLI dataset provides a strong foundation, expanding evaluation benchmarks to include additional languages and diverse NLP tasks (e.g., question answering, sentiment analysis) would allow for a more comprehensive assessment of cross-lingual prompting strategies.

Combining Prompting with Other Transfer Techniques: Combining prompting strategies like X-InSTA and In-CLT with other techniques, such as multitask learning and domain adaptation, may further enhance the ability of models to generalize knowledge across languages.

8. References

https://arxiv.org/pdf/2406.06608

https://arxiv.org/pdf/2210.03057

https://arxiv.org/pdf/2308.01223

https://arxiv.org/pdf/2305.04091

https://arxiv.org/pdf/2311.05232

https://arxiv.org/pdf/2310.14799

About the author

Naishadham Radha Sri Keerthi

Naishadham is an undergraduate student pursuing her B.Tech in Artificial Intelligence and Data Science trying to apply to universities for her master’s program in the same field.

The post An Empirical Comparison of LLM Prompting Strategies for Cross-Lingual Natural Language Inference appeared first on Exploratio Journal.

Abstract

Models’ lives may appear perfect—top fashion, after-parties, celebrity encounters, photoshoots, and designer brands are all things that come to mind when we think of the model lifestyle. But did you know that more than half of the individuals working in the fashion and modeling industry report using illicit drugs? In a period when drug usage is becoming increasingly prevalent among young women, the three main contributors to this are peer pressure, mental health challenges, and the use of drugs as a coping mechanism. These pressures, along with unique risk factors such as the need for the perfect figure, staying energized, standing out in a competitive field, and the “party culture,” can make models more prone to substance use. Some ways to help reduce drug use in the modeling industry include agency responsibility, accessible therapy, and drug testing.

Introduction

We see models everywhere: in magazines, billboards, commercials, and runway shows. Top-performing models’ lives can often appear perfect from an outside perspective with access to top fashion brands and items, extravagant after-parties, celebrity encounters, photoshoots, and global fame. However, behind the glamor lies a darker reality that is rarely discussed: the frequent drug use in the modeling industry. While drug use is a large-scale societal problem that can affect anyone regardless of their background, gender, occupation, or status, there are specific risk factors unique to models that increase their chances of substance abuse.

For young females specifically, the three main contributors to drug use are peer pressure, mental health, and use as a coping mechanism. A study conducted by research organization Nemours Kids Health found that peer pressure is one of the largest contributors to drug use among young people and teenagers. A study found that peer pressure caused 21% of teens to use drugs at least once (Thomas, 2023). Young adults are extremely likely to get peer pressured into drug use because of how hard it is for teenagers to say no and dread feeling left out of a certain social group.

Many young women may resort to drugs as a way of coping with things like anxiety, stress, and depression. These substances can provide temporary relief to these negatively perceived feelings and may feel like the only available escape. Drugs are also seen as ways to help adjust to new environments and situations when under stress, by providing a rare opportunity to relax.

In a clinical setting, drugs are prescribed to regulate mood by stabilizing dopamine and serotonin levels, leading to a temporary increase in a person’s mood. These pharmaceuticals can alter the way a person looks at themselves and the world around them, improving self-esteem and increasing their reported happiness. However, when used without appropriate recommendation from a professional, these drugs can be overused and create an unhealthy dependency.

Many young women start their substance abuse with the phrase-“I’ll try it once”, which is proven to be very unconvincing as the curiosity of trying something new often leads to a lifelong battle with addiction. According to a three-year longitudinal study, researchers recorded that among 12–15-year-olds, those who reported any tobacco curiosity at the baseline survey were most likely to later engage in use, which suggests curiosity may be one of the earliest markers of risk for later substance initiation (Khajeh Value, Zavar, Alidoust, & Pourandi (2013).

There is also a strong link between stress and substance abuse. Some people who work highly stressful jobs rely on drugs as a way of coping and escaping from the stress the jobs create for them. Research from OSHA reports that 83% of U.S. workers suffer from work-related stress and 54% say that work stress affects their home life. This is just for Americans in the workforce, but those who work in the modeling industry often experience much higher levels of work-related stress (Spooner & Hetherington, 2005).

Some of the most common contributors to work-related stress include limited control over one’s working schedule, fatigue, and longevity. Many of these factors are extremely present in the modeling industry. Models often do not understand what their upcoming work days may look like until they are living it, as jobs and shoots often get scheduled last minute. These jobs are often in close time frames with one another which leads to high levels of fatigue, and work weeks often include over 40 hours of diverse tasks.

Stress at work can lead to health-related issues that include anxiety, depression, and poor decision-making that leads to riskier behavior, such as substance abuse. Job industries prone to high stress levels such as the modeling industry are also subject to drug abuse. A survey from the American Addiction Center tells us that an average of 8.6% of workers across all occupations used drugs in the previous month because of work-related stress (Stein, 2019).

Industries with the highest levels of illicit drug use in the month prior to the survey were: Accommodation and food service (19.1%), Arts, entertainment, and recreation (13.7%), and Management (12.1%). With entertainment careers such as modeling, there is a high proportion of workers that turn to drug use in comparison to other industries (Free by the Sea, 2022).

As we saw in the previous point, there are many risk factors and contributors for young women that make them prone to drug use. Among all the other jobs and occupations, modeling seems to stand out in risk factors compared to those of different occupations.

More than half of individuals working in the fashion and model industry (54%) reported the use of illicit drugs. Illicit drugs are those not prescribed by a professional and taken and obtained illegally- and/or alcohol abuse at least once in their lives (Spooner & Hetherington, 2005). This rate is significantly higher than that reported by other industries, confirming that models represent a high-risk group for substance abuse disorders.

Drug use among fashion models is sometimes seen as a way to handle the stress and pressure of their jobs. In an industry where looks matter a lot, models might feel the need to stay thin and always fit into the expectations of society’s “perfect woman”. Some models turn to drugs like cocaine or amphetamines to help them stay thin and have more energy. Drugs can also provide them with a temporary sense of relief when they’re stressed, anxious about their appearance, or competing with other models.

Models have a very tight schedule which means that they have the constant need to be very energized without getting lots of rest. Lots of layovers and time changes when traveling, the need to stand out and look fresh at castings, and the need for energy in a very competitive field, are all reasons that models may take drugs as a coping mechanism.

A normal workday during events like New York Fashion Week frequently exceeds twelve hours. Models and stylists are expected to be at their best despite exhaustion and sleep deprivation. Stimulant drugs like cocaine can help cope with stress and fatigue, pull all-nighters, and be outstanding on a catwalk or in a studio for as long as it is required.

Another huge reason models may take drugs is to maintain their body weight or lose weight. In this profession, you must look a certain way to get jobs, and due to modeling being a very competitive field young women are willing to go to desperate measures to get the “perfect model figure.” Because of this, some models started using cocaine as not only a “stress reliever” but also a way to alter the body’s metabolism in a way that keeps a person skinny regardless of what they eat. Another stimulant, Noradrenaline also decreases your appetite and along with dopamine, breaks down stored fat, called lipolysis. When lipolysis occurs faster than usual, a person loses fat mass.

Appetite Suppressants, such as phentermine, diethylpropion, and liraglutide, work by making you feel less hungry and reducing your calorie intake. Stimulants, like amphetamines or methylphenidate, can increase metabolism and energy expenditure, thus leading to weight loss. Thyroid Hormones, such as levothyroxine, are known to increase metabolic rate. Fat Absorption Inhibitors, exemplified by drugs like orlistat, function by inhibiting the absorption of dietary fats.

However, research done by drug rehabilitation centers across the U.S. shows that the most common drugs in the fashion industry are cocaine and heroin, two highly dangerous and illegal drugs (12 Keys Rehab, 2016).

“Packs of cigarettes, daily colonics, laxatives, Phentermine diet pills, Adderall, prescription drugs that suppress the appetite… I’ve heard stories that some modeling agents encourage girls to do speed and cocaine to speed up their metabolism and eat less. And all kinds of injections are becoming more and more popular, from HCG injections that go with a 500-calorie diet plan to T3 thyroid injections that healthy models inject in an attempt to speed up their thyroid function, which results in a faster metabolism,” top fashion model Kira Dikhtyar quoted in an interview to Fox News in 2012.

Another specific risk factor unique to the modeling industry includes the idea of a “party culture.”

Party culture is defined as the “ideas, customs, and social behavior of a particular people or society,” according to the Oxford English Dictionary. Party culture follows this definition, except it would involve the social group of people who attend parties.

The party culture in the fashion industry can often influence fashion models to use drugs. In an environment where social events and gatherings are common and an easy way to boost social profile and find jobs, models may feel pressure to participate in late-night parties and after-show celebrations, occasions that are often fueled by alcohol and drugs, creating an atmosphere where substance use may seem normalized or even expected. The desire to fit in with peers and industry insiders, as well as the expectations of living up to the glamorous lifestyle associated with the fashion world, can contribute to models’ experimentation with drugs- experimentation that can lead to substance abuse.

Runway shows also often have a VIP area where models can go before, after, or during their show. These areas often have free drinks to help calm the models, which can lead to a reliance on drinks or other substances for stress release.

To contribute to the topic, I gathered data on drug use in the modeling industry by conducting an anonymous survey that questioned 55 female fashion models from 18-25 years old. The survey included questions about experiences with drug offers, pressures, and personal drug use resulting from the industry. The survey revealed that approximately 32% of models reported having been offered or pressured to use illicit drugs during their careers, and approximately 10% admitted to using illicit drugs at least once because of the industry.

A large part of this fight against substance abuse includes agency responsibility. A big way agencies can reduce the numbers of drug use is by drug testing their models- out of fear of losing jobs, models will be discouraged against drug usage. In a 2000 study by the National Institutes of Health, researchers discovered that the presence of drug testing and zero tolerance policies in the workplace significantly reduce the usage of substances in workers, as workers whose employers drug test were 0.57 times as likely to use marijuana in the past month than workers whose employers did not (NIH, 2000).

Another way agencies can help is by sending their models to therapy before the addiction even starts. This way, models would have the support of a therapist who teaches them what to do in certain situations and how to say no to pressure and possible desires to encourage a healthier lifestyle against stress. The National Institutes of Health also found in 2001 that interventions for work-related stress such as therapy have many benefits, including higher mood and lower levels of anxiety, which can help combat the demand for coping mechanisms such as illicit drug use (NIH, 2001).

While fashion models are in charge of their own decisions about drugs, they should know about available help and possible downfalls. They can be provided with a clear understanding of the risks and given the support to make choices that keep them healthy. If a young girl dreams of becoming a model due to the glamor and glory that is portrayed through the media, she should also be educated about the risks she is getting herself involved in. It is unlikely that her aspirations include unhealthy coping mechanisms. Models should understand that turning to drugs can also cause problems like addiction that may ultimately hurt their careers. They should be encouraged to prioritize their health over fitting in with what others are doing. Models can rewrite history and change society’s idea of the “Perfect American Model” from beauty standards to role models by showing they value staying healthy and making smart decisions. Fashion agencies and others in the industry should support models in staying safe and getting help if they need it, and it is also important for models to take control of their health and say no to things that could harm them to take back the modeling industry for the exciting environment it appears to be.

References

Abuse, N. I. on D. (n.d.). Increased drug availability is associated with increased use and overdose | National Institute on Drug Abuse (NIDA). Nida.nih.gov. Retrieved March 27, 2024, from https://nida.nih.gov/publications/research-reports/prescription-opioids-heroin/increased-d rug-availability-associated-increased-use-overdose#:~:text=NIDA.-

Amaro, H., Sanchez, M., Bautista, T., & Cox, R. (2021). Social vulnerabilities for substance use: Stressors, socially toxic environments, and discrimination and racism. Neuropharmacology, 188(188), 108518. https://doi.org/10.1016/j.neuropharm.2021.108518

Bogár, N., Dukay‐Szabó, S., Simon, D., Túry, F., & Pászthy, B. (2022). Frequency of disordered eating habits among fashion models. European Eating Disorders Review, 30(6). https://doi.org/10.1002/erv.2912

Canavan ’26, M. (2022). Call It What It Is: The Truth Behind the Modeling Industry. The Advocate. https://theacademyadvocate.com/13717/opinion/call-it-what-it-is-the-truth-behind-the-mo deling-industry/

Carpenter, C. S. (2007). Workplace Drug Testing and Worker Drug Use. Health Services Research, 42(2), 795–810. https://doi.org/10.1111/j.1475-6773.2006.00632.x
Drug Addiction in the Fashion Industry – 12 Keys Rehab. (2016, May 3). JourneyPure 12 Keys. https://www.12keysrehab.com/drug-addiction-in-the-fashion-industry/

Fixsen, A., Kossewska, M., & Bardey, A. (2022). I’m Skinny, I’m Worth More: Fashion Models’ Experiences of Aesthetic Labor and Its Impact on Body Image and Eating Behaviors. Qualitative Health Research, 33(1-2), 81–91. https://doi.org/10.1177/10497323221141629

High on the Runway: Drug Abuse in the Fashion Industry. (n.d.). Destinationhope.com. https://destinationhope.com/high-on-the-runway-drug-abuse-in-the-fashion-industry/
Ji, C. (2023, February 4). Heroin Chic and the Price of Beauty. Students, Faculty & Staff. https://m.georgetown.edu/welcome/student_news/detail?feed=student_news_1&id=04b0 d0c8-e01f-59b5-9665-95ee7f634601

Keyzers, A., Lee, S.-K., & Dworkin, J. (2020). Peer Pressure and Substance Use in Emerging Adulthood: A Latent Profile Analysis. Substance Use & Misuse, 55(10), 1–8. https://doi.org/10.1080/10826084.2020.1759642

Khajeh Value, M., Zavar, A., Alidoust, M., & Pourandi, R. (2013). The Relation of Self-Esteem and Illegal Drug Usage in High School Students. Iranian Red Crescent Medical Journal, 15(11). https://doi.org/10.5812/ircmj.7682

Lipari, R., & Van Horn, S. (2017, August 24). Children Living with Parents Who Have a Substance Use Disorder. Samhsa.gov. https://www.samhsa.gov/data/sites/default/files/report_3223/ShortReport-3223.html

Poulsen, T. (n.d.). The Dark Side of Our Fashion Industry: Drug Abuse. Anne of Carversville. https://anneofcarversville.com/style-photos/2019/2/7/the-dark-side-of-fashion-industry-is -drug-abuse

Santonastaso, P., Mondini, S., & Favaro, A. (2002). Are Fashion Models a Group at Risk for Eating Disorders and Substance Abuse? Psychotherapy and Psychosomatics, 71(3), 168–172. https://www.jstor.org/stable/48510101#:~:text=More%20than%20half%20of%20the

Sea, F. by the. (2022, January 26). Job-Related Stress and Addiction | Free by the Sea | Washington State. Free by the Sea. https://freebythesea.com/job-related-stress-and-addiction-how-work-burnout-can-lead-to- substance-abuse/

Spooner, C., & Hetherington, K. (2005). SOCIAL DETERMINANTS OF DRUG USE. https://ndarc.med.unsw.edu.au/sites/default/files/ndarc/resources/TR.228.pdf
Stein, S., August 21, Msnl. U., & 2019. (n.d.). Occupations Linked to Drug and Alcohol Abuse. American Addiction Centers. https://americanaddictioncenters.org/occupational-stress-influences

The benefits of interventions for work-related stress. (2001). American Journal of Public Health, 91(2), 270–276. https://doi.org/10.2105/ajph.91.2.270

Thomas, A. (2023). Peer Pressure with Substance Use. https://www.thenationalcouncil.org/wp-content/uploads/2023/07/23.06.29_Getting-Candi d-Peer-Pressure-with-Substance-Use-Blog-2.pdf

Wade, N. E., Palmer, C. E., Gonzalez, M. R., Wallace, A. L., Infante, M. A., Tapert, S. F., Jacobus, J., & Bagot, K. S. (2021). Risk factors associated with curiosity about alcohol use in the ABCD cohort. Alcohol, 92, 11–19. https://doi.org/10.1016/j.alcohol.2021.01.002

Why Are Some People More Prone to Addiction? (2018, July 3). Priory. https://www.lifeworkscommunity.com/blog/why-are-some-people-more-prone-to-addicti on#:~:text=Children%20of%20

About the author

Darya Ilnytska

Darya Ilnytska is currently a senior at Poway High School. She models on the side and is currently signed with agencies in Milan, LA, OC, and SD. Through her experiences in the modeling industry and her love for sociology and psychology, Daria has become passionate about raising awareness on important issues like mental health, social pressures, and substance use. Her research aims to shed light on these challenges and explore solutions to help improve the well-being of models and others in the industry.

The post Drug Use in The Modeling Industry appeared first on Exploratio Journal.

Abstract

In a world where artificial intelligence is beginning to shape critical life decisions, can we trust that the algorithms guiding these choices are unbiased? This paper investigates the implications of algorithmic bias in hiring processes, emphasizing the dual role of artificial intelligence (AI) as both a transformative tool for recruitment and a potential perpetrator of discrimination. It begins with a review of current hiring practices and then identifies key factors contributing to algorithmic bias, including data quality issues, algorithmic opacity, and the influence of proxy variables. Notable cases where biases have emerged, including Amazon’s recruitment algorithm, which favored male candidates due to biased training data, are carefully examined. The paper outlines various strategies for mitigating algorithmic bias while acknowledging their limitations, such as data augmentation, vector space correction, and blind hiring. Furthermore, the research extends its analysis beyond hiring, exploring the manifestations of algorithmic bias in facial recognition technology, predictive policing, and healthcare, thus illustrating the broader societal implications. Conclusively, the paper advocates for creating strong frameworks and legislation to promote more openness and responsibility when using algorithms, underscoring society’s moral obligation to ensure technology serves all communities equitably in an increasingly automated world.

Key Terms

Algorithmic bias:‬‭ systematic and repeatable errors‬‭ in a computer system that create unfair outcomes

Applicant Tracking System (ATS):‬‭ a software system‬‭ that helps organizations manage the hiring and recruiting process

Artificial Intelligence (AI):‬‭ computer software systems that are capable of performing tasks traditionally associated with human intelligence

Proxy Variables:‬‭ a variable that serves as a substitute‬‭ for the variable of interest that cannot be measured directly

Target Variable:‬‭ a feature of a dataset needing to‬‭ be understood more clearly

1. Introduction

Today’s businesses focus more on finding the right employees to maintain their competitive edge. To achieve this, many companies are turning to artificial intelligence (AI) embedded within applicant tracking systems (see Key Terms) to streamline hiring processes, enhance efficiency, and reduce workloads. However, while AI offers significant advantages, it can also unintentionally perpetuate discrimination in hiring. This occurs when biased algorithms or data lead to unfair treatment of certain candidates (see Figure 1). Resources like the Implicit Association Test, available at‬‭ https://implicit.harvard.edu/implicit/takeatest.html,‬‭ can be valuable tools to help individuals recognize and explore their biases. Understanding the causes of discrimination in hiring is not just crucial; it’s a responsibility we all share for developing fairer and more inclusive employment practices.

This paper will first explore the current job hiring process, some of the key issues that lead to algorithmic bias in the status quo, and real-world examples of hiring bias. Possible solutions to mitigating these biases will also be evaluated, acknowledging current limitations but highlighting ongoing advancements and future potential. Finally, general applications of algorithmic decision-making across various fields will be explored, demonstrating the breadth of these tools and the significance of addressing biases to create a fairer, more inclusive future for all.

2. Navigating AI-Driven Hiring

2.1 Applicant Tracking Systems

Applicant tracking systems, otherwise known as ATS, have become increasingly common in hiring practices. Websites embed this software system to help recruiters filter candidates throughout the hiring process, improving applicant sourcing. When candidates apply to a job posting by sharing their resumes, an ATS will scan candidates based on qualifying questions that satisfy the company’s standards (see Figure 2). AI plays a crucial role in this process, scanning resumes based on specific parameters regarding skills, qualifications, experiences, etc. It takes over the tedious human task of shortlisting candidate resumes, but algorithms may have encoded societal stereotypes found in the data they were trained with‬‭ (Frissen et al., 2023)‬‭ .

2.2 Three Factors Contributing to Algorithmic Bias

Algorithmic bias is shaped by several key factors, beginning with data quality issues. These issues arise when the training data used to develop algorithms is biased, incomplete, or reflective of historical inequalities. For instance, if data is collected from an organization that has historically disproportionately hired more white employees than Black employees, the algorithm might associate good performance with being white. This does not mean that hiring only Black employees would fix the bias; instead, it’s crucial to improve the diversity and balance of the data to reduce the bias. The urgency of this need cannot be overstated. Proxy variables (see Key Terms) can also embed systemic biases within algorithms even when direct indicators of discrimination, such as race and gender, are not explicitly included. For example, zip code can be a proxy variable for race because it strongly correlates with neighborhood segregation‬‭ (Fountain, 2022)‬‭ .

Another significant factor contributing to algorithmic bias is algorithmic opacity.‬‭ This term refers to the potential lack of understanding of an algorithm due to its complexity. Often, the lack of transparency in an algorithm makes it hard for human users to interpret its internal processes (Sadek et al., 2024)‬‭ . Many algorithms operate as “black boxes,” meaning their decision-making processes are not easily understandable or interpretable by those affected by their outcomes. If the logic behind decisions is unclear, it becomes challenging to identify or correct biases, ensure fairness, or hold the creators accountable for discriminatory outcomes.

Algorithms also rely on correlating variables with a target variable (see Key Terms) to predict outcomes.‬‭ For example, if a tech company was looking to hire a software developer proficient in a specific programming language such as Python, the target variable could be “proficiency in Python.” The recruitment algorithm would then categorize candidates into groups based on their coding skills, such as “expert in Python,” “basic Python knowledge,” “no Python experience,” etc. This allows the company to narrow the pool of candidates to those who match the technical expertise required for the job.

However, a key problem with target variables is how the output is defined will influence the result. For example, suppose non-technical skills (soft skills) are considered important for an organization and are part of the target variable. In that case, women may gain an advantage compared to algorithms that do not consider such skills. Additionally, the measure of employees’ performance is also based on the subjective assessment by their managers. Factors such as the employee-manager relationship, personal biases, or differing expectations within the team environment can contribute to the algorithmic evaluation. As a result, algorithms trained on this data may unintentionally reflect biases present in the workplace, resulting in unfair outcomes where high-performing individuals who may fit the requirements might be overlooked or undervalued by the algorithm.

2.3 Current Hiring Practices

AI has played a growing role in hiring and transforming how companies identify, evaluate, and recruit talents.‬‭ Early AI-driven hiring systems were‬‭ used to make the hiring process more efficient but lacked sophisticated methodologies. For example, Resumix was founded in 1988 and served as a resume parsing tool. ATS made its debut in the 1990s with job posting sites such as CareerBuild. By the early 2000s, talent assessment tools like eSkill and SkillSurvey used AI to automate pre-employment testing and background/skill checks. The 2010s saw the rise of AI-powered video interviewing software, with platforms like HireVue utilizing machine learning algorithms. Natural language processing (NLP), a subfield of AI that uses machine learning to understand spoken and written human language, is used to analyze speech patterns, word choice, and language structure during video interviews to assess candidates. Sentiment analysis provides insights into candidates’ emotions and engagement, adding another layer to talent evaluation.

However, AI’s integration into hiring has led to concerns about bias. A Microsoft research study in 2019 highlighted significant biases with AI algorithms embedded in the data on which they are trained. Researchers found that language models like Word2Vec, a machine learning technique that uses NLP to obtain vector representations of words, could produce biased associations between specific demographic groups and stereotypical terms. For example, their investigation led them to observe outputs such as a man is to a woman as a computer programmer is to a homemaker‬‭ (Chiu, n.d.)‬‭ . These biases pose risks in applications like resume screening, where hidden associations could unintentionally favor or disadvantage certain groups.

In 2018, Amazon’s AI-driven algorithm was also found to be biased. When AI systems are trained on historical data, they often reflect the existing biases within that data. If a company’s past hiring practices were skewed, those biases could be unintentionally embedded in the AI’s decision-making process. When Amazon attempted to automate its recruitment process in 2018, it used an algorithm trained on the previous 10 years of resumes. The dataset of resumes consisted primarily of male applicants over ten years, causing the algorithm to favor male language patterns and resulting in discrimination against female candidates‬‭ (Dastin, 2022)‬‭ . Words such as “executed” and “captured” were commonly found on male engineers’ resumes and were more favored by the technology. The system also downgraded resumes that featured the word “women’s,” such as in “women’s chess club captain.” This example highlights the potential risks of relying on AI in hiring, as biased training data can perpetuate discrimination and undermine diversity efforts.

While Amazon’s algorithm faced scrutiny for discriminating against women, it also prompted broader questions about how personal data is utilized in hiring. Companies increasingly rely on AI algorithms to sift through resumes, assess candidates, and predict job performance (see Figure 3). This reliance raises crucial considerations regarding the transparency and accountability of data usage in recruitment processes. Globally, regulatory approaches to AI in hiring are evolving. The European Union has taken the lead in establishing AI policy with its AI Act, which aims to ensure that AI hiring systems follow strict privacy rules. It mandates transparency in AI decision-making and imposes requirements for assessing the impact of AI on employment outcomes‬‭ (Sadek et al., 2024)‬‭ .

In the U.S., the rules for AI are less organized and vary from place to place. With general agreement that there need to be AI policies, the question becomes who will make the rules? A commonly shared point of view is that frameworks must ensure that company technologies do not cause harm and that they are held accountable for their impacts. Furthermore, policies need to advocate for greater transparency, including how AI systems work and the data they use. Actors in the AI space must adopt principles that promote responsible AI use, as articulated in the White House’s Blueprint for an AI Bill of Rights‬‭ (‬‭ The Three Challenges of AI Regulation‬‭ , n.d.)‬‭ .

As AI technology advances, the need for coherent frameworks will become increasingly important to ensure fairness and accountability in employment practices.

3. Solutions and Limitations of Mitigating Algorithmic Bias in Hiring

As algorithms play a more significant role in hiring, techniques like data augmentation, vector space correction, and blind hiring offer valuable ways to enhance fairness and inclusivity (see Figure 4). While each method brings its own limitations, they represent significant strides toward reducing bias in AI-driven recruitment.

3.1 Data Augmentation

Definition:‬‭ Data augmentation involves using existing‬‭ training data and modifying it to create new instances that enhance machine learning model training. This technique helps address the issue of insufficient data by artificially increasing the volume, quality, and diversity of training data‬‭ (Mumuni & Mumuni, 2022)‬‭ .

Mechanism:‬‭ Common data augmentation methods for images include rotating, flipping, or cropping. For text data or data used in hiring practices, data augmentation techniques include synonym replacement, paraphrasing, and numerically variating data (see Figure 5).

Advantages:‬‭ Augmenting data increases the diversity‬‭ of the training dataset, allowing models to generalize unseen data better. Expanding a dataset also helps reduce overfitting, where a model memorizes the training data rather than learning its underlying patterns. Data augmentation allows a model to recognize patterns, increasing its ability to handle real-world variations.

Limitations:‬‭ However, augmented data may not always reflect real-world scenarios, potentially leading to overfitting if the modifications are not representative. Extreme data augmentation can introduce excess and unimportant information, deteriorating a model’s quality‬‭ (Walidamamou, 2023)‬‭ .

3.2 Vector Space Correction

Definition:‬‭ The vector space is a mathematical framework in which data points, such as words and images, are represented as vectors in a multidimensional space. Vector space correction helps mitigate biases by equalizing the distance between the protected attributes (such as race or gender) and the biased concept‬‭ (Albaroudi et al., 2024)‬‭ . See Figure 4 for a simplified visual demonstration.

Mechanism:‬‭ The process involves adjusting the positions‬‭ of vectors to reduce biases. For example, if a vector model leans towards associating white people with better skills and qualifications, the vector space correction technique will associate the same skills and qualifications with Black people.

Advantages:‬‭ This technique helps create a more balanced‬‭ representation of different groups, which can reduce the impact of biased data on model predictions.

Limitations:‬‭ Vector space correction can cause semantic drift, where adjustments in the vector space may unintentionally change the meanings and relationships of the data points. This can lead to inaccurate predictions and misinterpretations of ideas, making it harder for algorithms to accurately reflect real-world scenarios. Another limitation of this approach is that biases related to more than one attribute are hard to correct because many factors need to be considered before rearranging the vector space.

3.3 Blind Hiring

Definition:‬‭ Blind hiring is a recruitment strategy‬‭ that aims to eliminate bias by removing personal information from decision-making systems. Personal information such as names, zip codes, and health records can sometimes be indicators of social class, gender, age, or racial background.

Mechanism‬‭ : This technique focuses on evaluating candidates‬‭ based solely on their skills and qualifications, removing the chances of being unconsciously influenced in hiring decisions.

Advantages‬‭ : One advantage of blind hiring is that‬‭ it promotes diversity by allowing candidates from varied backgrounds to compete on an equal footing. Ultimately, this practice can lead to a more inclusive workplace culture overall.

Limitations‬‭ : While blind hiring practices aim to eliminate visible identifiers, such as names and genders, they do not fully address the underlying gender, racial, or social biases in the hiring process. This is because specific keywords can still influence perceptions and decisions, as they carry implicit biases that favor one group over another, regardless of the removal of direct identifiers (see Figure 6). For example, masculine traits typically include characteristics like confidence and competitiveness, whereas feminine traits often encompass emotional qualities like warmth, supportiveness, and collaboration‬‭ (Albaroudi et al., 2024)‬‭ .

4. Addressing Algorithmic Bias in Broader Decision-Making Systems

Beyond job hiring, cases of algorithmic bias appear in various fields, including facial recognition technologies, predictive policing, and healthcare (see Figure 8). The following sections will explore how these biases manifest in these areas and examine their implications.

4.1 Facial Recognition Technology

Facial recognition technologies (FRT) are used to identify faces in static or moving images. The accuracy of an FRT depends upon the quality of the image it assesses and the makeup of the algorithm itself. FRTs are popular in authentication processes, police work, and medical diagnosis. However, many FRTs have been found to exhibit algorithmic bias, leading to disparities in accuracy based on race, gender, and other demographic factors. FRTs first capture the details of an image, identifying if it is human or not. A person’s face is broken down into key features, such as the distance between the eyes and the shape of the cheekbones. This information is translated into a faceprint, uniquely given to each individual. The faceprint is compared to images in a database to find a possible match (see Figure 7). False positives and false negatives are possible. A false positive is misreading an image as a match when it is not, whereas a false negative fails to match the face. Challenges such as the quality of available images, lighting, and facial expressions can affect the accuracy of FRTs.

Studies show that FRTs are often more accurate for lighter-skinned males but tend to misidentify women and individuals with darker skin tones at significantly higher rates. The Amazon Rekognition System resulted in higher accuracy for white and black men than white women (93%) and dark-skinned women (68.6%). The U.S. National Institute of Standards and Technology (NIST) also found higher rates of false positives for Asian and African American faces in comparison to Caucasian faces using the FBI’s database of 1.6 million domestic mugshots‬‭ (Fountain, 2022)‬‭ .

Further diversifying datasets to train from more racial and ethnic groups will help mitigate algorithmic bias and promote fairer outcomes. However, establishing and adhering to rigorous standards is also essential to improve the quality and accountability of this technology.

4.2 Predictive Policing

Predictive policing is a law enforcement technique that uses data and algorithms to predict where and when crimes will occur. The goal is to use this information to prevent crime, but it has emerged as a controversial approach to law enforcement. One of the early implementations, CompStat in New York City during the 1990s, employed visual tools like pin maps to display crime data by frequency and location‬‭ (Fountain, 2022)‬‭ . However, while CompStat aimed to promote efficiency and accountability, it also contributed to problematic practices such as “stop and frisk,” which is the practice of stopping individuals for questioning, sometimes without reasonable suspicion. This practice disproportionately targeted racial minorities, and research has shown that these practices can lead to lasting psychological harm.

The practice has raised significant concerns about algorithmic bias and its implications for marginalized communities. Some algorithms may produce biased results that lead to over-policing or police being repeatedly deployed to neighborhoods based on skewed data of specific neighborhoods. Some municipal governments have implemented executive orders banning the use of predictive policing software. However, while bans are effective in the short run, they are not a substitute for legislative action. At the root, a larger focus on addressing algorithmic bias is crucial to ensure that predictive policing does not aggravate existing inequalities in the criminal justice system.

4.3 Healthcare Algorithms

In the healthcare industry, algorithms have also been shown to lead to racial bias. A recent study found bias in an algorithm that generated individual-level medical risk scores affecting 200 million people. The algorithm identifies patients for “high-risk care management” but relies on healthcare costs as a proxy for illness, leading to biased outcomes. The algorithm assumed that those with higher medical costs were sicker. However, Black patients, despite having higher levels of illness than their White counterparts with the same risk score, tend to generate lower healthcare costs due to limited access to care and implicit biases in care. Additionally, Black patients tend to spend more on emergency visits and dialysis costs rather than inpatient surgeries and outpatient specialist costs, which cost more‬‭ (Fountain, 2022)‬‭ .

This research emphasized that biases often arise from flawed labels reflecting structural inequalities. Addressing these biases through improved algorithm design and iterative testing can lead to fairer outcomes, opening pathways for more equitable healthcare solutions.

5. Conclusion

Algorithmic bias poses significant challenges across various sectors of the economy, often disproportionately affecting marginalized communities. As hiring practices evolve, companies increasingly rely on AI to make recruitment more efficient, yet this shift has also introduced unintended barriers to diversity, equity, and inclusion (DEI). Organizations aiming to improve the DEI culture must prioritize transparency, accountability, and ethical considerations in their hiring algorithm designs. While technical solutions are essential, they often are not enough on their own. This shows how crucial it is to have strong laws to tackle algorithmic bias. Passing effective legislation is a moral responsibility to ensure that technology benefits everyone. The future of mitigating algorithmic bias depends on global collaboration. By passing thoughtful laws, we can build a technology environment that helps everyone, creating a fairer society that supports communities in a world that relies more and more on automation.

6. References

Albaroudi, E., Mansouri, T., & Alameer, A. (2024). A Comprehensive Review of AI Techniques for Addressing Algorithmic Bias in Job Hiring.‬‭ AI‬‭ ,‬‭ 5‬‭ (1), Article 1. https://doi.org/10.3390/ai5010019

Bîgu, D., & Cernea, M.-V . (2019).‬‭ Algorithmic Bias‬‭ in Current Hiring Practices: An Ethical Examination‬‭ .

Chiu, R. (n.d.).‬‭ Can We Fix AI Hiring Bias? | Policy‬‭ Commons‬‭ . Retrieved October 24, 2024, from‬‭ https://policycommons.net/artifacts/1320212/can-we-fix-ai-hiring-bias/1923502/‬‭ .

Dastin, J. (2022). Amazon Scraps Secret AI Recruiting Tool that Showed Bias against Women*. In K. Martin,‬‭ Ethics of Data and Analytics‬‭ (1st ed.,‬‭ pp. 296–299). Auerbach Publications. https://doi.org/10.1007/s00146-022-01574-0‬‭ .

Fountain, J. E. (2022). The moon, the ghetto and artificial intelligence: Reducing systemic racism in computational algorithms.‬‭ Government Information Quarterly‬‭ ,‬‭ 39‬‭ (2), 101645. https://doi.org/10.1016/j.giq.2021.101645‬‭ .

Frissen, R., Adebayo, K. J., & Nanda, R. (2023). A machine learning approach to recognize bias and discrimination in job advertisements.‬‭ AI & SOCIETY‬‭ ,‬‭ 38‬‭ (2), 1025–1038. https://doi.org/10.1007/s00146-022-01574-0‬‭ .

Lomibao, L. (2020, December 13). Factsheet: Facial Recognition Technology (FRT).‬‭ Stop LAPD Spying Coalition‬‭ .‬‭ https://stoplapdspying.org/facial-recognition-factsheet/‬‭ .

Mumuni, A., & Mumuni, F. (2022). Data augmentation: A comprehensive survey of modern approaches.‬‭ Array‬‭ ,‬‭ 16‬‭ , 100258.‬‭ https://doi.org/10.1016/j.array.2022.100258‬‭ .

Sadek, T., Stanley, K. D., Smith, G., Marcinek, K., Cormarie, P., & Gunashekar, S. (2024). Artificial Intelligence Impacts on Privacy Law‬‭ . RAND‬‭ Corporation. https://www.rand.org/pubs/research_reports/RRA3243-2.html‬‭ .

Sen, S. (2023, September 22). Applicant Tracking System: The Ultimate Guide to Smart Hiring. Asanify. https://asanify.com/blog/human-resources/applicant-tracking-system-the-ultimate-guide-t o-smart-hiring/‬‭ .

The Ultimate Guide to AI in recruiting [2024]‬‭ . Joveo.‬‭ (2024, October 4). https://www.joveo.com/the-ultimate-guide-to-ai-in-recruiting/‬‭ .

The Three Challenges of AI Regulation‬‭ . (n.d.).‬‭ Brookings. Retrieved October 19, 2024, from https://www.brookings.edu/articles/the-three-challenges-of-ai-regulation/‬‭ .‬

About the author

Sanaa Gada

Sanaa is a passionate high school senior looking to study computer science in college, with a focus on developing technology that bridges the gap for marginalized communities. She enjoys giving back to the community via projects that support the rising generation such as leading science experiments and teaching programming to underprivileged youth. In her free time, Sanaa enjoys baking, dancing, hiking, and babysitting.

The post Fair or Flawed? How Algorithmic Bias is Redefining Recruitment and Inclusion appeared first on Exploratio Journal.