AI roots out three key predictors of terrorism support

A new study suggests that individuals who justify terrorism tend to share a distinct worldview characterized by the normalization of violence, moral flexibility, and anti-democratic sentiments. Using a machine learning analysis of data from over 90,000 people across 65 countries, the research provides evidence that support for terrorism is embedded in a broader set of attitudes rather than being an isolated belief. The findings were published in the journal Aggressive Behavior.

Public discourse often links terrorism to specific religious or political ideologies, yet extremist violence emerges from a wide range of backgrounds, including white supremacist and anti-government movements. This suggests that the psychological and social factors underlying radicalization may be more general than commonly assumed.

Previous research has identified many potential risk and protective factors for aggression, but these studies have often been limited to smaller, nationally specific samples. The current research aimed to apply a comprehensive, data-driven approach on a global scale to identify which attitudes and beliefs are most consistently associated with the justification of terrorism.

“The study was motivated by evidence that extremist ideologies are evolving in many parts of the world. For example, in recent years, pro-violence right-wing movements, white supremacist groups, and other forms of radicalization within established democracies have gained strength and attracted increasing numbers of followers,” said study author Mohsen Joshanloo, an associate professor at Keimyung University and honorary principal fellow at the Centre for Wellbeing Science at the University of Melbourne.

“These developments suggest that some traditional assumptions about who justifies violence, including terrorism, may no longer hold. Public discourse has not fully kept pace with these changes and requires revision. The aim of this study was to identify patterns in a global dataset and provide an updated, evidence-based understanding of the factors associated with justifying terrorism.”

To conduct the analysis, the researcher utilized data from the seventh wave of the World Values Survey, a large-scale project that gathers information on the beliefs and values of people around the globe. The final sample included 91,659 participants from 65 nations. The study’s outcome of interest was a single question that asked respondents to rate the justifiability of “terrorism as a political, ideological or religious means” on a 10-point scale.

A machine learning algorithm known as Random Forest was employed to analyze the predictive power of 360 different variables from the survey. This method is well-suited for identifying complex patterns in large datasets without being constrained by predefined assumptions. The algorithm builds hundreds of decision trees to determine which factors are the most important predictors of an outcome, in this case, the justification of terrorism.

The final model, which included 271 of the most relevant predictors, was able to explain approximately 64% of the variation in people’s attitudes toward terrorism justification. The analysis revealed that a small number of predictors were especially powerful. These top factors fell into three distinct but related domains.

The first domain was a general normalization of violence. Individuals who were more likely to justify terrorism also tended to believe that other forms of aggression are acceptable. This included the justification of political violence, violence against other people, a man beating his wife, and parents beating their children. This pattern suggests that support for terrorism is not an isolated belief but part of a wider cognitive framework where aggression is viewed as a legitimate tool for resolving conflicts or achieving goals.

A second major pattern involved what the study describes as moral flexibility and a tendency to violate rules. People who justified terrorism were also more likely to approve of dishonest behaviors such as accepting a bribe, stealing property, cheating on taxes, and claiming government benefits one is not entitled to.

These individuals may operate with a moral code that permits breaking established social and ethical norms when it appears personally or ideologically useful. This mindset aligns with psychological concepts of moral disengagement, where individuals cognitively separate their actions from their moral principles, allowing them to engage in unethical behavior without self-censure.

The third key domain was a preference for religious and political authoritarianism. Respondents who justified terrorism were more likely to support a system where religious authorities interpret laws and were less likely to see free elections as an essential characteristic of democracy. They also tended to view a system of governance based on religious law, without political parties or elections, more favorably.

This combination of attitudes points to a skepticism toward democratic institutions and a preference for strong, centralized, and non-pluralistic forms of rule. The authoritarian preference appeared to be focused on political structures, as these same individuals showed more permissive attitudes toward certain personal behaviors like casual sex and homosexuality.

“Those who consider terrorism acceptable often share a worldview in which violence is normalized, moral rules are flexible, and democratic principles are devalued,” Joshanloo told PsyPost. “This worldview is not confined to any single group, religion, or region. Anyone can hold it or adopt it, regardless of location or identity. Outdated and simplistic explanations ignore the fact that pro-violence views are not limited to any particular demographic. To effectively reduce support for terrorism (and violence more broadly), we must address the underlying attitudes that legitimize violence everywhere, without assuming that one’s own group or oneself is immune to such beliefs.”

“Strategies to reduce violence should be guided by current evidence, not outdated stereotypes. For example, strengthening democratic institutions, addressing their weaknesses, and building community trust and cooperation are crucial steps, even though these measures are far more challenging than relying on easy but ineffective assumptions and prejudices.”

The findings can also be interpreted through the lens of the “Dark Triad” of personality traits: Machiavellianism, psychopathy, and narcissism. Although these traits were not measured directly, the observed patterns align with them conceptually.

“For example, endorsing bribery, fraud, and other rule- breaking behaviors for personal gain may reflect Machiavellian tendencies, that is, strategic, self-serving disregard for norms,” Joshanloo explained. “The normalization of violence and lack of empathy align with psychopathic traits, which involve callousness and instrumental use of aggression.”

“Narcissism may be reflected in the preference for authoritarian governance and rejection of democratic principles, signaling an inflated sense of ideological entitlement and disdain for pluralism. Taken together, these connections suggest that support for terrorism is not only ideological but may also stem from deeper psychological dispositions that erode empathy and moral constraints.”

While the study offers a robust global perspective, it has some limitations. The analysis was restricted to the variables available in the World Values Survey, meaning other potentially important factors, such as exposure to extremist propaganda or group grievances, were not included. Because the data was cross-sectional, it identifies associations but cannot establish causality.

Future research could build on these findings by incorporating a wider range of variables, using more detailed psychological measures, and employing longitudinal designs to better understand how these attitudes develop over time.

“It is important to note that this study does not examine engagement in terrorism, but rather the extent to which individuals view terrorism as justifiable,” Joshanloo noted. “The results should be understood as general global patterns, offering a broad picture of what tends to predict justification of terrorism globally. They are not meant to describe any single country, region, group, or context in detail.

“A factor that appears unimportant at the global level might be highly significant in a specific local setting. Therefore, these findings should serve as a backdrop for more nuanced local research rather than a substitute for it. The value of global insights is that they reveal common psychological and ideological tendencies that transcend borders, helping us understand what is broadly associated with terrorism justification worldwide. This big-picture perspective provides a reference point for interpreting local findings and designing strategies that combine global insights with context-specific solutions.”

The study, “Who Considers Terrorism Justifiable? A Machine Learning Analysis Across 65 Countries,” was authored by Mohsen Joshanloo.