China's new set of regulations on labeling artificial intelligence generated content (AIGC) will officially come into effect on Monday.
The regulations, jointly released by the Cyberspace Administration of China (CAC) and three other departments, stipulate that all text, images, audio, video, virtual scenes and other contents generated by artificial intelligence should be explicitly or implicitly labeled on dissemination platforms.
Platforms should conduct reviews before content release to identify the markers of AIGC, and add risk notices to unidentified or suspected AIGC, so as to block the transmission of false information, according to the regulations.
"There are three levels of review in the regulations. If the platforms detect implicit markers from the metadata of contents, they can then label them as AI-generated. However, if the platforms merely infer through some algorithms that the contents are generated by AI, the contents should be labeled as suspected AIGC. The regulations serve as basic guidelines for dissemination platforms to handle content with a multi-tier reviewing and labeling system," said Zhang Jiyu, executive director of the Renmin Law and Technology Institute under Renmin University of China. As technological limitations can still result in the false labeling of some original content as AI-generated at the moment, Zhang said mechanisms should be built to deal with such situations to better protect the interest of content creators.
"n the future, when relevant platforms all employ the multi-tier labeling system for AIGC, they should also have multi-tier handling systems for the contents. For instance, when the content is clearly identifiable as AI-generated through implicit markers, stricter measures should be applied. However, for content that should be labeled as potentially AI-generated, more accessible channels for user feedback should be provided, and users should not face severe penalties," said Zhang.
First, the platforms should disclose algorithmic principles and misjudgment sample data to enhance transparency. Second, they should establish appeal channels allowing users to submit evidence for review requests, thereby ensuring a balance between regulation and creative freedom, Zhang added.
China to implement new regulations on labeling AI-generated content
