Social media, algorithms, echo chambers, online gaming, and generative artificial intelligence are creating an environment that helps amplify extremist ideas.
A new study by TRENDS Research & Advisory finds that digital transformation has reshaped the environment of extremism and radicalization. It warns that social media, algorithms, echo chambers, online gaming, and generative artificial intelligence (AI) are creating an environment that amplifies extremist ideas, accelerates their spread, and expands opportunities for polarization and recruitment.
The study, prepared by Maryam Al Junaibi, a researcher in the Terrorism and Extremism Studies Sector at TRENDS Research & Advisory, explained that extremist propaganda has shifted, in less than two decades, from publications, tapes, and recorded sermons to an interactive digital space that enables extremist actors to measure audience responses and continuously adjust their messaging according to levels of engagement. The trend is no longer limited to jihadist organizations, but has extended to right-wing extremism, anti-women groups, and transnational conspiracy theory networks.
The study noted that algorithms do not create extremists out of nothing, but can accelerate a process rooted in social, psychological, and political factors by repeatedly exposing users to polarizing content and reducing their exposure to opposing viewpoints. At the same time, echo chambers reinforce feelings of belonging and polarization.
It further explained that constructing digital identity is a key stage in this process, as ideas can evolve from consumed content into collective identities people feel compelled to defend. Communication may then shift from open platforms to private conversations and encrypted apps, allowing direct relationships to form with individuals more susceptible to influence.
The study warned that online gaming, chat platforms, and livestreaming services are increasingly important spaces for polarization and recruitment. It cited the Gaming the System study, which surveyed 1,128 players across five countries and found that 51 percent of respondents had encountered extremist rhetoric or narratives while playing multiplayer games. Players under the age of 18 were more likely than adults to encounter certain forms of extremist content.
The study also examined emerging challenges associated with generative AI, noting its potential to facilitate the production and customization of extremist propaganda. It highlighted documented cases of AI-generated images used in campaigns linked to extremism and conspiracy theories, while stressing that extremist groups’ exploitation of these technologies remains relatively experimental.
The study emphasized that the transition from exposure to extremist content to the adoption of extremist ideas or violence is not an inevitable process, but may involve several stages, beginning with exposure and repetition, followed by belonging and mobilization, and potentially leading, in a limited number of cases, to escalation and violence. In this context, it cited studies by RAND Europe and the U.S. National Institute of Justice, which indicate that the internet is generally neither the sole nor the decisive factor in an individual’s radicalization pathway.
The study called for an integrated approach to countering digital extremism, including greater transparency and accountability in recommendation systems, enhanced media and digital literacy among young people, and the development of tools to detect extremist rhetoric in gaming environments and live voice chats. It also called for adopting the “safety by design” principle and maintaining human oversight in generative AI applications.
The study concluded that digital platforms do not create extremism from scratch but can amplify its underlying drivers and accelerate its growth. The more complex challenge lies in their ability to turn extremist content into social relationships, collective identities, and networks of belonging, making the response a shared responsibility that requires integrating preventive, social, technological, and legal approaches.
