Artificial intelligence could significantly accelerate the process of radicalization by amplifying psychological vulnerabilities and exposing users to increasingly extreme content online. Published in the journal Personality and Social Psychology Review, the research combines psychological theories of radicalization with modern AI technologies such as recommendation algorithms, generative AI systems, and automated bot networks. The researchers argue that AI-driven systems can intensify feelings of anger, isolation, and injustice in ways that may gradually push vulnerable individuals toward extremist beliefs.
The study outlines a four-stage process through which radicalization can develop. First, algorithms expose users to polarizing content, even if they are not actively searching for it. Next, repeated recommendations and personalization reinforce those beliefs, creating echo chambers that strengthen ideological views. Over time, online communities—and potentially AI-generated companions or bots—can foster a sense of belonging and identity around extremist narratives. In rare but dangerous cases, this progression may eventually contribute to real-world violent extremism.
Researchers also highlight how generative AI introduces new risks beyond traditional recommendation systems. Unlike older algorithms that mainly decide what users see, generative AI can actively create persuasive propaganda, fake communities, deepfakes, and highly personalized extremist messaging at massive scale. The study warns that AI-generated interactions may make it increasingly difficult for users to distinguish between genuine human influence and automated manipulation. Because these systems are optimized for engagement, they may unintentionally exploit emotional vulnerabilities such as loneliness, identity insecurity, or social marginalization.
At the same time, the researchers emphasize that AI alone does not “create” radicalization from nothing. Psychological vulnerability, social conditions, and existing grievances still play a central role. However, AI technologies may dramatically increase the speed, scale, and personalization of extremist influence compared to previous online systems. The findings add to growing global concerns about the societal impact of AI and the need for stronger safeguards, transparency, and accountability in algorithmic platforms that shape public discourse and online behavior.
