China's internet regulator has launched the second phase of its "Clear and Bright" campaign targeting chaos in AI apps — and the cleanup is already massive. Regulators say they've wiped out more than 5.61 million pieces of rule-breaking content, taken action against more than 49,000 accounts, and dealt with more than 2,400 unlawful websites and apps. The official focus: AI-generated false information, violent and vulgar content, impersonation, and material that harms minors.
On the moderation side, Douyin, Kuaishou, Weibo, Tencent, Baidu, Bilibili, Xiaohongshu, Zhihu, Douban, Taobao, and other platforms are strengthening their multimodal recognition tools, expanding face databases, voiceprint databases, and violation sample libraries, and cracking down harder on violators. The regulator says they've published 46 governance notices during the campaign.
Meanwhile, AI platforms including Doubao, Yuanbao, Qwen, and ERNIE Bot are tightening the gate: stricter review of raw training corpora, stricter limits on rule-breaking output, and stronger requirements to label AI-generated or synthetic content. App stores operated by Huawei, Xiaomi, OPPO, and vivo are also tightening developer onboarding and review, with spot checks and re-tests on apps that are already available.
During this phase, cyberspace authorities also called out typical violations: using AI to remix classic literature, fabricate disaster and trending-topic information, swap faces and voices to impersonate public figures, mass-produce violent or vulgar content, generate content that harms minors, and use AI account-management software to run coordinated "water army" troll operations. Offending AI products and apps can expect warnings, fines, delisting, or closure under Chinese law.
Source: https://www.cac.gov.cn/2026-09/02/c_1790099041364574.htm?utm_source=chatgpt.com