Key Takeaways
1. Mandatory AI Content Labeling: All AI-generated text, images, audio, and videos must be clearly labeled to indicate their origin.
2. Platform Compliance: Major platforms like WeChat, Douyin, Weibo, and Xiaohongshu have implemented specific requirements for labeling AI-generated content with strict enforcement.
3. Penalties for Misuse: The Cyberspace Administration of China (CAC) has established penalties for those who misuse AI to spread misinformation or manipulate public opinion.
4. Focus on Misinformation and Protection: The law aims to combat misinformation, protect copyrights, and enhance online safety, especially for young users.
5. Global Trend: China is among the first major markets to enforce such labeling laws, with similar regulations emerging in other regions worldwide.
China has put a new law into place that mandates social media platforms to tag content made by AI. This law went into effect on Monday, and major platforms began using their new compliance tools on the same day. This action from Beijing is part of the CAC’s “Qinglang” initiative for 2025, aimed at fighting misinformation and increasing control over AI.
New Labeling Requirements
All text, images, audio, and videos produced by AI must have clear labels indicating they are AI-generated. Additionally, they should include hidden identifiers like digital watermarks or metadata. This law was created by the Cyberspace Administration of China in conjunction with industry experts, public security, and broadcasting authorities. It was announced in March and has now been put into action.
Platform-Specific Changes
WeChat, which has about 1.4 billion active users each month, now requires content creators to announce if their posts are AI-generated. They strictly prohibit any alteration of these labels. Douyin, known as TikTok in many countries, which boasts around 766.5 million monthly active users, mandates visible tags on all AI-generated content and checks sources via metadata. Weibo has introduced a feature for reporting “unlabeled AI content,” while Xiaohongshu, also called rednote, will add labels if users forget to do so. Many platforms are also advising users to be careful with content that lacks labels and may remove posts that are incorrectly labeled.
Penalties and Focus Areas
The CAC has set penalties for those who misuse AI to spread false information or sway public opinions, particularly targeting paid online commentators. This law seeks to address misinformation, copyright issues, and online scams, especially focusing on misleading marketing in short-video applications and aiming to enhance protections for young users. While this promotes transparency online, there are still chances that some may evade these labeling requirements.
China stands as one of the first big markets to enforce this mandatory labeling for AI-generated content, with similar regulations popping up in other regions globally. The Internet Engineering Task Force has suggested an AI metadata header, and devices like the Google Pixel 10 are starting to include C2PA credentials. It’s worth noting that users can still occasionally bypass these safety measures.
Source:
Link
