China has removed more than 5.61 million pieces of AI-generated or AI-related content as regulators widen a crackdown on how artificial intelligence is used online.
The Cyberspace Administration of China (CAC) announced Wednesday that the second phase of its “Qinglang” campaign against AI application abuse has scrubbed more than 5.61 million pieces of illegal or non-compliant content, penalized over 49,000 accounts, and taken action against more than 2,400 websites and apps.
The four-month campaign, which launched in April, targets AI-generated content used to spread false information, violent or vulgar material, impersonate real people, or harm minors, according to the CAC.
The CAC’s report described AI remixes of classics like “Romance of the Three Kingdoms” and “Journey to the West” that twist characters into clickbait plots, content the regulator has branded “digital slop.”
Other flagged cases included AI-generated fake disaster footage used to scare or mislead elderly users, deepfaked videos impersonating public figures for profit, and content sexualizing or endangering minors. The CAC also called out AI hosting software used to automate social media accounts, comments and repetitive posts, describing it as an AI-powered evolution of China’s so-called internet “water armies.”
Platforms and chatbot makers on the hook
Major platforms including Douyin, Kuaishou, Weibo, Tencent, Baidu, Bilibili, Xiaohongshu, Zhihu, Douban and Taobao have expanded face, voiceprint and prohibited-content sample libraries used to improve automated detection and issued 46 governance notices combined, per the CAC.
Leading chatbot developers — ByteDance’s Doubao, Tencent’s Yuanbao, Alibaba’s Qwen and Baidu’s Ernie Bot — have tightened reviews of training data and restricted prohibited outputs, according to the CAC’s report. The agency said those platforms “strictly control the review of original corpus content, strictly limit the output of illegal content, strengthen the implementation of AI-generated synthetic content identification requirements, and prevent the generation and dissemination of illegal and irregular information from the source.”
App stores run by Huawei, Xiaomi, Oppo and Vivo have also tightened developer vetting, the CAC said. The crackdown builds on China’s AI labeling rules, which took effect Sept. 1, 2025, and require AI-generated and synthetic content to carry explicit or metadata-based identifiers in specified circumstances. Major platforms have reportedly labeled more than 150 billion pieces of AI-generated or synthetic content.
More must-read AI coverage
- SS&C Intralinks DealCentre AI vs. Datasite: Which platform is built for the future of dealmaking?
- SS&C Intralinks FundCentre AI vs. Juniper Square: Which platform better supports modern private markets fund managers?
- Why Data, Not Models, Determines AI Success
- The Rise of the AI-Native Factory: How Physical AI Is Transforming Manufacturing
The bottom line
The interesting part of this campaign is that China is moving beyond simply identifying AI content after it appears online.
The CAC is also pushing controls before generation and distribution, including training-data review, output restrictions and app-store screening. For AI companies operating in China, compliance increasingly extends beyond removing prohibited posts after publication to controls across training data, model outputs, content labeling and app distribution.
For consumers, stronger detection could mean fewer convincing scams, impersonation videos and misleading AI posts. But stricter filtering can also raise concerns about false positives and how platforms distinguish harmful synthetic content from legitimate creative uses.
The CAC said it will continue monitoring emerging AI risks and take further measures to curb misuse while supporting the orderly development of AI applications.