The Cyberspace Administration of China has launched the second phase of its campaign targeting misconduct in AI applications, requiring platforms such as DouBao and YuanBao to strictly restrict unauthorized content.

The Cyberspace Administration of China has launched the second phase of its campaign targeting misconduct in AI applications, requiring platforms such as DouBao and YuanBao to strictly restrict unauthorized content.

2026-09-02 06:27View Original

According to a notice from the Cyberspace Administration of China (CAC), the CAC has launched the second phase of the "Clear and Bright – Rectifying Abuses in AI Applications" campaign, focusing on addressing issues such as AI-generated false information, violent and vulgar content, impersonation and imitation of others, and infringement of minors' rights and interests. To date, the campaign has removed over 56.1 million pieces of illegal and non-compliant information, suspended more than 49,000 accounts, and dealt with over 2,400 websites and applications found in violation. In terms of AI content governance, platforms including Douyin, Kuaishou, Weibo, Tencent, Baidu, Bilibili, Xiaohongshu, Zhihu, Douban, and Taobao have further strengthened their multimodal recognition capabilities, dynamically expanded facial image databases, voiceprint libraries, and repositories of prohibited content samples, and intensified enforcement against rule-violating accounts and information, issuing a total of 46 governance notices.

Meanwhile, AI platforms such as Doutai, Yuanyuan, Qwen, and Wenxin Yiyan have tightened the review of original training data, strictly restricted the output of prohibited content, and reinforced requirements for labeling AI-generated and synthesized content. Huawei, Xiaomi, OPPO, vivo, and other app stores have enhanced developer onboarding and app listing reviews, conducting spot checks and retesting of existing applications. Furthermore, the cyberspace administration has disclosed several typical cases, including AI-driven "remixing" of classic literary works, generation of false disaster and social hotspot information, AI-based face and voice spoofing of public figures, mass production of violent and vulgar content, use of AI to generate information infringing upon minors’ rights and interests, and employing AI-powered "account management" tools for online mob activities. For non-compliant AI products and applications, relevant authorities will take legal measures including warnings, fines, removal from app stores, or shutdowns.

Disclaimer: Contains third-party opinions, does not constitute financial advice

Recommended Reading

Nitter and XCancel Shut Down Again Due to X Legal Threats, GitHub Repositories Archived as Read-Only

7 days ago
Nitter and XCancel Shut Down Again Due to X Legal Threats, GitHub Repositories Archived as Read-Only

Anthropic Opens Near-Internal Access to External Security Teams, CEO Says Model Capability Development Should Be Slowed Down

10 days ago
Anthropic Opens Near-Internal Access to External Security Teams, CEO Says Model Capability Development Should Be Slowed Down

Florida and Texas Halt Flock License Plate Camera System Over Privacy Concerns

21 days ago
Florida and Texas Halt Flock License Plate Camera System Over Privacy Concerns

Flock Safety CEO Calls for Compromise on Privacy and Security, Company Faces Accusations of Surveillance Abuse

08-23
Flock Safety CEO Calls for Compromise on Privacy and Security, Company Faces Accusations of Surveillance Abuse

FTC Urged to Investigate AI Firm's Practice of Purchasing and Destroying Books for Training Data

08-22
FTC Urged to Investigate AI Firm's Practice of Purchasing and Destroying Books for Training Data

OpenAI Launches New Security Measures Following Hugging Face Incident: Enhanced Monitoring, Network Isolation, and 30-Minute Alert System

08-19
OpenAI Launches New Security Measures Following Hugging Face Incident: Enhanced Monitoring, Network Isolation, and 30-Minute Alert System

U.S. data centers could consume two to four times more water by 2028 than in 2023, prompting multiple regions to tighten factory construction approvals

08-06
U.S. data centers could consume two to four times more water by 2028 than in 2023, prompting multiple regions to tighten factory construction approvals