Beijing Introduces New Rules to Label AI-Generated Content | Festina Lente - Your leading source of AI news | Turtles AI

China has unveiled a draft regulation to require digital platforms to identify AI-generated content, with visible metadata and logos. This approach aims to improve transparency online and maintain strict control over the provenance of published content.

Key points:

• Mandatory labeling for all AI-generated content. 
• Use of visible logos and embedded metadata in files. 
• Platforms responsible for verifying suspicious content. 
• Requirement to retain user activity logs for six months.

China’s internet regulator has proposed a new regulatory framework that, if confirmed, could mark a further step towards stricter management of AI-generated content on digital platforms. The plan, currently in draft stage, requires that all files created or modified using AI tools be clearly identified with a visible label and accompanied by specific metadata. This label must be present not only on the surface, but also embedded in the file’s metadata to ensure full traceability. In the case of audio content, the proposal calls for a voice prompt to inform listeners of the artificial nature of the content both at the start and end of playback, with the option to insert it during the broadcast if necessary. Video players will also have to comply with similar rules, displaying clear warnings at relevant points in the video. Netizens who use AI tools to create content will be required to report such creations, and service providers will have to keep detailed records of their activities for a period of six months. Without metadata, digital service providers will be required to analyze content and label it if there is any suspicion that it was generated by AI. Another aspect of the plan is for Chinese internet platforms to pool expertise to collectively improve their systems for detecting artificial content. Any violation of the new rules could lead to significant consequences for the platforms and individuals involved. While the proposal is still open for public comment until October, it is unlikely to be substantially changed, as drafts submitted by Chinese authorities typically tend to be implemented without major changes. The proposal underscores Beijing’s commitment to maintaining tight control over online activity, ensuring that user-generated content can always be traced back to identifiable individuals, further tightening its surveillance of the local digital landscape. This level of scrutiny is intended to ensure that any criticism of the government or the Communist Party is quickly identified and dealt with.

The draft represents a new measure by Beijing to tightly regulate and monitor AI-generated content, amid the growing use of AI technologies.