The law issued in March requires explicit and implicit labeling of text, images, audio, video and other virtual content generated by AI. Explicit markings must be clearly visible to the user, but implicit identifiers such as digital watermarks must be embedded in the metadata.
The new regulations reflect an increasing scrutiny of AI in Beijing, as concerns increase over misinformation, copyright infringement and online fraud.
It also coincides with a broader push to tighten AI surveillance, becoming a key focus for CAC's 2025 Qinglang, Or a clear and bright campaign – an annual initiative aimed at cleaning China's cyberspace.
WeChat, known as the mainland Weixin, said that content creators must voluntarily declare content generated for all AI at the time of publication. Regarding content that is not flagged, WeChat said it reminds users to “take their own judgment” online. According to Tencent's latest financial report, WeChat adds over 1.4 billion monthly active users worldwide.
