AI-Generated Disturbing Videos Harm Children's Mental Health in China
AI-generated disturbing videos featuring bizarre character combinations and violent or sexual content have become popular among Chinese schoolchildren through short video platforms. Parents report children become obsessed with the content, and psychological experts warn that prolonged exposure may damage children's attention capacity and distort their understanding of traditional culture.
What happened
AI-generated videos known as 'ε€ε½ε±±ζ΅·η»' (foreign Shanhaijing) and 'AIιͺε Έ' (AI obscene content) have become popular among Chinese elementary school children. These videos feature bizarre, AI-generated character combinations with violent and bloody content, including images of mutilation, eye-gouging, and disturbing remakes of classic characters such as Ultraman depicted as pregnant or Pinocchio hitting someone. Users can easily create such videos by submitting basic character images, text descriptions, and instructions to common AI video websites.
Short video platforms continuously recommend disturbing content to children despite user attempts to report videos or mark them as disliked. According to a former video company employee, hundreds of video accounts under a major platform have persistently promoted such 'obscene animations' by applying AI generation technology to classic animated characters. The videos use bright colors to attract children and employ algorithmic recommendation to create a distribution feedback loop.
Reported impact
Parents report that children have become obsessed with these videos, repeatedly speaking mysterious words from the content and showing signs of being 'possessed.' A child in Guangdong province was frequently exposed to disturbing animated videos after watching legitimate cartoons and animation. According to psychological consultants, prolonged exposure to such high-stimulation, low-density 'electronic junk food' may cause fragmentation of adolescents' attention mechanisms and weaken the brain's capacity for deep reading and logical thinking.
Children reportedly develop distorted views of traditional culture. Long-term viewing of such 'cybernetic Shanhaijing' may cause children to associate the classical text with crude, distorted content rather than its actual cultural significance. The phenomena has extended to commercial products, with merchandise featuring these characters sold on e-commerce platforms, reportedly creating social pressure and hidden competition among children.
Index analysis
The content bypasses platform moderation systems by using keyword substitution to generate inappropriate 'second-dimension' style images with sexual undertones and other harmful elements, including depictions of child pregnancy. AI modification packages such visual content as 'creative innovation' while degrading classic literature and traditional culture, potentially embedding the false belief in young minds that classics can be arbitrarily deconstructed.
Governance controls
Platforms lack effective mechanisms to prevent the continuous promotion of harmful content to minors even after users report videos or indicate disinterest. The current moderation and recommendation systems fail to stop the algorithmic promotion loop that amplifies exposure to disturbing content.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.