China has introduced new measures for labeling AI-generated synthetic content, effective September 1, 2025. Emphasizing transparency and user protection, these rules require both explicit and implicit labeling for various forms of content, including text and video, to ensure audiences can identify the origins of AI-generated materials. Companies providing such content must comply with updated regulations in order to protect user rights and interests.
In a significant move to regulate artificial intelligence, four Chinese government agencies—including the Cyberspace Administration of China—unveiled new guidelines on March 14 for labeling AI-generated synthetic content. These rules, effective from September 1, 2025, aim to enhance the healthy development of AI while ensuring that citizens and organizations are protected, as outlined in the introduction of the Measures.
The measures, enacted on March 7, categorize AI-generated content into several forms: text, images, audio, video, and virtual scenes. Explicit and implicit labeling are both included. Explicit labels come in the form of visible notifications that users can easily recognize, like text or sound, while implicit labels are woven into the content’s metadata and are less noticeable.
Applicable to internet-based content providers using AI for content generation, these measures also stipulate adherence to existing regulations surrounding recommendation algorithms and deep synthesis technology. High-risk AI-generated content—think intelligent dialogues and speech synthesis—will need to feature explicit notifications that alert users to its origins.
Specifically, the guidelines detail how to implement these labels: they may include text alerts or audio signals at key points in videos, images, and interactive scenarios. Thorough deliberation is given for where these labels should be placed to mitigate user confusion. Furthermore, implicit labels will capture essential metadata like the service provider’s name and a reference number of the content.
A notable provision in Article 7 mandates that online platforms conducting reviews for applications that utilize generative AI must require explanations from service providers regarding their content labeling. Moreover, content-generating companies will need to update their internal policies ahead of the compliance deadline.
Importantly, Article 10 emphasizes that users must actively declare any synthetic content when publishing it online, using the labeling functions offered by the respective platform. Any manipulation or deceptive practices around these labels—whether by organizations or individuals—are strictly prohibited.
As China pushes for leadership within the realm of AI technology, this framework signals a commitment to safeguarding users’ interests and promoting transparency in content dissemination. Interestingly, the juggling of innovation and responsibility paints a complex landscape for the future of AI in the country.
These new measures reflect China’s dual ambition of leading in artificial intelligence while also ensuring protections for its citizens against potential misinformation. With explicit and implicit labeling requirements, the initiative seeks to bolster transparency and accountability in AI-generated content. As we push toward the September 2025 deadline, all eyes will be on the effectiveness of these regulations in shaping the future of digital content.
Original Source: medium.datadriveninvestor.com
