GNET warns of extremist AI videos on TikTok

Researchers from the Global Network on Extremism and Technology (GNET) have warned that AI-generated videos of anthropomorphic fruits and vegetables committing violent acts are spreading extremist agendas on TikTok. These clips, which use a “cute” or “Pixar-like” aesthetic, are reportedly bypassing the platform’s moderation policies regarding extreme violence.

The report indicates that the content includes messages linked to the Islamic State (IS), extreme right “accelerationism,” and “nihilistic violence extremism.” Specific examples include a video of a carrot causing a potato to “explode” and another featuring a “Carrot Khalifah” using slogans reminiscent of IS propaganda. Some accounts also reportedly reference the neo-Nazi Atomwaffen Division.

Experts suggest the danger lies not only in the ideological messaging but in the “normalisation of cruelty” and the erosion of emotional boundaries through constant exposure to graphic AI-generated gore. This trend of “AI slops”—low-quality, mass-produced digital content—is seen as a method of progressively desensitizing young users to violence.