China’s AI content labeling law is now in force, and the country’s biggest social platforms switched on new compliance tools the same day. The rule is part of the Cyberspace Administration of China’s 2025 “Qinglang” campaign targeting misinformation, online manipulation, and abuse of generative AI.
What the law requires
– Every AI-generated text, image, audio, and video must carry a clear, visible label stating it is AI-generated.
– Content must also include hidden identifiers, such as digital watermarks or metadata, to support source verification and enforcement.
– The regulation was drafted by the CAC with input from industry, public security, and broadcasting regulators. It was issued in March and has now taken effect.
How major platforms are complying
– WeChat (Weixin, ~1.4 billion MAU): Creators must declare AI-generated content. The platform prohibits tampering with labels.
– Douyin (~766.5 million MAU): Requires visible labels on every AI-generated post and verifies sources through metadata.
– Weibo: Introduced an “unlabeled AI content” reporting option so users can flag suspect posts.
– Xiaohongshu (RED): Applies labels itself if users fail to do so.
– Several services now prompt users to treat unflagged content cautiously, encourage reporting, and reserve the right to remove mislabeled or deceptive posts.
Enforcement priorities and penalties
The CAC signals penalties for using AI to spread misinformation or manipulate public opinion, with heightened scrutiny on paid online commentators. Beyond misinformation, authorities cite goals of curbing copyright infringement and online fraud, tightening oversight of deceptive marketing on short‑video apps, and improving protection for minors. Transparency is the focus, though officials acknowledge that label evasion remains a risk.
Global context
China is among the first major markets to mandate AI-content labels and watermarks. Similar efforts are emerging worldwide, including proposals for standardized AI metadata headers at the Internet Engineering Task Force and expanding support for C2PA credentials on consumer devices such as the Google Pixel 10. Even with these safeguards, determined users may still find ways around controls, underscoring the need for continual vigilance.
What creators and brands should do now
– Always apply visible “AI-generated” labels and preserve watermark/metadata signals throughout your workflow.
– Avoid editing steps that strip metadata; audit export and compression settings.
– Use platform-native tools for labeling and source verification to reduce takedown risks.
– Be transparent in captions and descriptions, especially for ads and sponsored posts.
– Train teams on the new requirements and document compliance for audits.
– Monitor platform policy updates and user reports, and respond quickly to disputes.
Bottom line
The new labeling mandate sets a strict, platform-wide baseline for AI transparency in China. With major services already enforcing the rules, creators, marketers, and media teams should update their processes immediately to protect reach, trust, and compliance in an increasingly regulated AI landscape.
