China enforces mandatory labels for AI-generated content across social media platforms

Artificial Intelligence

Chinese users logging into apps like WeChat, Douyin (TikTok’s Chinese twin), Weibo and RedNote are now seeing labels attached to posts that include AI-generated text, images, audio or video. According to South China Morning Post, the new rules require not just visible tags but also hidden identifiers such as watermarks embedded in metadata.

Mandatory AI labels may also reshape user habits globally. US tech firms have started experimenting with similar systems — Google’s Pixel 10 phones, for instance, automatically embed C2PA content credentials in images. China’s move, however, shows what such measures look like when backed by law and strict enforcement.

WeChat has told its users they must proactively add these labels when posting AI-generated content. The company also made it clear that people are banned from tampering with or removing labels that the platform applies automatically. It further warned against using AI to spread false information, infringe copyrights, or engage in illegal activities.

Douyin has issued similar instructions, stressing that its systems can detect the origins of content through metadata even if users skip labelling. Weibo, meanwhile, has added a feature that allows users to report unlabelled AI content, creating an enforcement layer powered by its community.

The regulations were drafted earlier this year by four government agencies, including the Cyberspace Administration of China (CAC), the Ministry of Public Security, and the National Radio and Television Administration. The CAC has already been running a campaign to tighten oversight of AI apps and services, according to the report.