The Cyberspace Administration of China, the Ministry of Industry and Information Technology, the Ministry of Public Security, and the National Radio and Television Administration have jointly issued new regulations introducing mandatory labeling for AI-generated content. These measures will take effect on September 1, 2025, and require platforms to clearly mark content created by AI systems.
The new rules build on a range of existing laws and regulations, including the Cybersecurity Law, as well as regulations governing algorithmic recommendations and deep synthesis technologies. Their primary goal is to ensure that the public can easily recognize when texts, images, videos, or audio clips have been generated by AI.
Under the new measures, AI service providers such as DeepSeek or Midjourney must visibly label AI-generated content. Clear warnings must appear either at the beginning, middle, or end of the content. In addition to these visible labels, providers must embed an implicit label—a digital watermark in the file header containing technical information such as the source, code, and the company’s identification number.
Platforms like TikTok and Weibo, which distribute online content, are also obligated to verify whether uploaded content has been AI-generated and label it accordingly. Users uploading AI-generated materials must self-identify such content using the labeling features provided by the platforms. Manipulating, removing, or altering these labels—or offering tools that enable such actions—is strictly prohibited.
However, some aspects remain unclear. It has not yet been determined whether content that is edited by humans after being generated by AI must still be labeled as AI-generated. Moreover, the effectiveness of these measures could be undermined if enforcement mechanisms remain vague. Currently, penalties are only broadly linked to overarching laws such as the Civil Code and Competition Law, which could result in inconsistent enforcement.
The new labeling requirements could have profound impacts on existing IP strategies. Companies may need to revise their protection mechanisms to safeguard both human-created and AI-assisted works, potentially leading to significant changes in copyright policies, licensing models, and enforcement strategies.
Source: State Council of the People’s Republic of China
China Intellectual Property Blog 