AI content must be labeled in China from September

Considering the technical advances in artificial intelligence, China will introduce a far-reaching obligation to label AI content on September 1, 2025. This will also place app store operators such as Apple and Google under obligation. The Communist Party's censorship authority wants to put a stop to the misuse of AI and the spread of misinformation.

The new labeling requirement applies to all types of media such as text, audio, image and video. According to an FAQ from the responsible ministries, AI products must be labeled as such both explicitly and implicitly. This means that, for example, an image must include a visual reference that identifies it as AI-generated. At the same time, however, the metadata must also be marked accordingly. The methods must be disclosed via the user agreements of the respective services.

App stores have a duty

China already has regulations on the handling of AI products. The new regulations are intended to clarify these. China is not the only country to take technological progress into account. The European Union has also created a legal framework for dealing with AI with the AI Act.

China also holds marketplaces for apps responsible. Before publication, they must check whether the app in question contains AI-generating functions and whether it meets the prescribed labeling requirements.

Transition period planned

Under certain conditions, the new regulation also provides for exemptions from labeling. At the same time, it is forbidden to delete or conceal labels. However, it is also punishable to falsely label man-made content as an AI product.

It remains unclear what penalties will be imposed in the event of an infringement. Initially, a six-month transition period is to begin in September so that companies have sufficient time to make the necessary technical adjustments.

(mki)