China’s AI-Labeling Measures and Mandatory National Standards Take Effect September 1

The article reports that China's cyberspace regulator, together with the industry and information technology ministry, the public security ministry and the national broadcasting authority, issued rules on labeling AI-generated synthetic content on March 14, 2025. The rules, called the Measures, are due to apply from September 1. A cybersecurity standardization committee and the State Administration for Market Regulation also finalized national standards on the labeling method, which likewise start on September 1.

The regulator describes the Measures as a way to end the misuse of generative AI and the spread of false information. The article adds that two figures at the recently concluded "Two Sessions" meetings, Lei Jun of Xiaomi and the actor Jin Dong, each proposed laws or regulations for AI-generated content.

Providers must put visible labels on AI-generated or synthesized text, images, audio, video and virtual scenes, and add hidden labels to the metadata of the files. If a service allows downloading, copying or exporting of such content, the output files must keep the visible labels. Platforms that review apps before listing them must ask developers whether generative AI services are offered and check the labeling materials. People who publish synthetic content through online information services must declare it and use the platform's labeling functions.

Violations are to be handled by the authorities for internet information, telecommunications, public security and broadcasting. The national standards define how visible labels should look, for instance an "AI" marker in text, as a superscript, or through sound and rhythm, and what metadata the hidden labels contain. Businesses have until September 1 to study the documents and build labeling tools.