Anti-extremism researchers are warning that AI-generated videos featuring violent fruits and vegetables could push children toward radicalization.
The videos initially appear similar to other viral “brain rot” clips, with colorful, human-like foods starring in bizarre stories about relationships, betrayal, and family drama.
However, researchers from the Global Network on Extremism and Technology found that some accounts have transformed the trend into violent videos designed to shock viewers.
The GNET report said the videos often use bright colors, cartoon characters, and animation reminiscent of children’s movies to disguise disturbing material.
Researchers found examples on TikTok in which animated fruits and vegetables used weapons against other food characters. Some clips also appeared to borrow imagery, slogans, and themes associated with extremist movements and criminal organizations.
Researchers warn violent AI fruit clips can become ‘attention traps’ #
According to the report, AI tools have made it cheap and easy for creators to produce large numbers of videos with little human involvement.
These clips are then optimized for endless scrolling, with simple stories, constant surprises, and repetitive visuals keeping viewers engaged.
Researchers warned that the content could “make violence seem funny, normal or even exciting to emulate,” particularly among younger viewers.
One disturbing clip shows a knife-wielding potato attacking a carrot tied to a tree. Another features a ‘Carrot Khalifah’ telling an army that “the Carrot State will rise again,” while other videos show a carrot attacking a potato with a saw and a hammer.
The report also suggested that comment sections can act as gateways to more troubling online communities. Users may explain the hidden references behind videos or direct others toward increasingly extreme content.
Related #
While the researchers stopped short of claiming every violent food video is deliberately created as propaganda, they warned that repeated exposure can normalize cruelty and weaken viewers’ emotional response to disturbing imagery. The authors called on TikTok and AI companies to introduce stronger detection systems that can recognize violent content even when it is hidden behind seemingly harmless cartoon characters. TikTok prohibits extreme violence on its platform.
The warning follows previous concerns about AI YouTube videos aimed at children that contained dangerous lessons, including incorrect road safety advice and unsafe food recommendations.