Radicalization is a multifaceted process driven by various interconnected factors. AI researcher Mijke van den Hurk has delved into whether artificial intelligence can help clarify this complex phenomenon. On June 29, she celebrated her Ph.D. defense, graduating with honors.
The catalyst for her research came in the spring of 2019 when van den Hurk was vacationing in Sri Lanka during a devastating terrorist attack that resulted in over 250 fatalities from suicide bombings targeting churches and luxury hotels. The assaults were attributed to a group linked to the radical Islamist organization IS.
"It was shocking to witness the aftermath. The pervasive fear and uncertainty were palpable. The attackers, who were affluent and well-educated—including some women—didn’t conform to typical stereotypes of terrorists. It prompted me to reflect on what could lead individuals to commit such acts," van den Hurk recounted.
This incident spurred her to explore the topic further through books, podcasts, and documentaries, revealing the complexity behind why individuals might join extremist groups. Factors such as the need for recognition, social status, peer pressure, and environmental influences all play a role.
Given the serious ramifications of terrorism, understanding radicalization is of utmost importance. As a recent graduate of the MSc Artificial Intelligence program and a new data scientist with the police, van den Hurk sought to leverage AI in this field. Through a partnership between Utrecht University and the Dutch police, she could dedicate herself to this research within the National Police Lab AI, stating, "Collaborating with law enforcement allows us to share knowledge and enhance our understanding of this intricate topic."
Van den Hurk emphasizes that, despite extensive research, no singular theory fully explains the process of radicalization. "The attraction to a violent ideology is fundamentally different from the act of committing violence itself. There are numerous variables at play, making targeted intervention challenging."
She investigated the potential of agent-based modeling (ABM) as a means to uncover insights. ABM allows researchers to visualize human behavior by simulating interactions among various entities over time. "Imagine a virtual space populated by individuals with diverse personality traits who respond to one another and their environments based on specific behavioral rules. By analyzing these interactions, we can observe the development of radicalization."
Van den Hurk created a model that merges various theories about the radicalization process, illustrating how a person's agency fluctuates through different social and psychological influences. Her simulations revealed how factors like social pressure, feelings of exclusion, identity formation, and exposure to extremist ideologies can either promote or counteract radicalization.
The model indicated that radicalization often arises from usual social dynamics that only coalesce under particular conditions, leading to extreme behaviors. "The trajectory of radicalization isn’t fixed by individual traits but by the most impactful mechanisms at that moment. For instance, we found that certain small groups became radicalized as a response to a dominant majority, but only when they lacked alternative ways to fulfill their social needs."
Van den Hurk's research holds promise for deeper understanding through a potential interactive dashboard that could allow users to manipulate various parameters to explore different scenarios of radicalization. While her research has not definitively answered the question that arose during her Sri Lankan vacation, it has significantly improved her comprehension of radicalization, both as a societal issue and through her personal reflections.
