Skip to content
Submit incident
AII-1001-57609
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.

AI-Generated 'Foreign Shanhaijing' Videos Target Chinese Minors with Violent and Sexual Content

AI-generated videos called 'Foreign Shanhaijing' became viral among Chinese minors in 2025, featuring absurd AI-combined creatures with nonsensical names. The videos contain violent and sexual content that distracted children from studies and led them to mistake the foreign content for China's classical Shanhaijing text, according to CCTV reporting and Beijing's Cyberspace Administration.

24 December 2025 · Beijing Internet Office of Cyberspace Administration · Harmful and abusive content
24 December 2025
AII-1001-576093 sources ↓
Quick facts
Category
Harmful and abusive content
Country
China
Source
凤凰网(凤凰新媒体)
Date
24 December 2025
Drafted by
Raidu Incident Detection

What happened

AI-generated short videos known as 'Foreign Shanhaijing' became widely popular among Chinese minors starting in 2025. The videos originate from overseas 'brain-rot meme' videos that use AI technology to randomly combine animal, object and human characteristics to create absurd creature forms, paired with distorted audio and invented story backgrounds.

These videos were reposted from overseas platforms to domestic Chinese short-video platforms beginning in early 2025, sparking large-scale secondary creation and transmission. Related merchandise including toys, stationery, comic books and small games based on these AI-generated images have been mass-produced and sold on online shopping platforms and in school neighbourhood shops.

Reported impact

Children have become intensely absorbed in memorizing nonsensical names from the videos, with some learning them faster than school curriculum. Parents report children do not know China's actual Shanhaijing classical text and mistakenly believe Shanhaijing is foreign.

Videos contain AI-animated characters depicting violent scenes including mass killings, beatings and cutting, and some related games embed soft pornography content. The content risks normalizing violence and sexual material to minors and disrupting their studies.

Index analysis

AI image generation tools require only a few keywords to produce animation content in seconds, lowering technical barriers to creation. Platform recommendation algorithms amplify reach by continuously promoting videos to users who view even briefly, creating a cycle of continued exposure.

The videos use exaggerated, sensational and high-impact visuals designed to capture minors' attention. According to Beijing's Cyberspace Administration, primary and secondary school students treat knowledge of the images as social currency, creating peer competition.

Governance controls

Law experts note that under the Regulation on Network Information Content Ecology Governance, platforms are responsible for content review, filtering and removal, and should prioritize healthy and beneficial content in recommendation weights.

Beijing's Cyberspace Administration identified the technical threshold decline and algorithmic promotion as drivers of the spread. Manual human review of reported content involves delays measured in minutes to days, while algorithm-driven recommendations are instantaneous.

Classification

Category
Harmful and abusive content
Industry
Media, social platforms and marketing
Country
China
Harm status
Actual harm
Intent
Unintentional
Timing
Unknown
Harm types
Psychological, Public interest
Harmed
Children
Failure modes
Hallucination, Misidentification
AI systems
AI image generation tools

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

凤凰网(凤凰新媒体)
央视曝光"外国山海经视频",内容涉暴力软色情
24 December 2025
驱动之家
动物人类随机拼接!央视曝光外国山海经:当心AI魔改怪兽带偏孩子
24 December 2025
华龙网
这些爆火视频需警惕!"AI拼接怪兽"正在带偏你的孩子
24 December 2025
Apr 2026AI-Generated Pornography Distribution Chain Exposed in ChinaConfirmedHarmful and abusive content中华网军事频道中华网科技公司Jun 2025AI deepfake tools sold for nude images in China criminal networkConfirmedHarmful and abusive contentJNRTP - Rádio Televisão Portuguesaexecutivedigest.sapo.ptMar 2025Sichuan authorities probe AI outbound calling robots in harassment scandalUnder investigationHarmful and abusive content东方财富网新浪财经guancha.cn+3Mar 2025CCTV Exposes AI-Powered Harassment Call Scandal at XiaokrConfirmedHarmful and abusive contentfinance.eastmoney.comnews.china.comfinance.sina.com.cnSep 2026AI Digital Humans Pose as Medical Experts to Advertise Products in ChinaWell documentedMisinformation and synthetic mediacj.sina.com.cnnews.bjd.com.cn中华网科技公司+5