AI Content Now Must Have a ‘Label’
From September 1, 2025, China has implemented strict regulations requiring all content generated by artificial intelligence (AI) to have mandatory ‘labels’. This includes text, images, audio, video, and even virtual reality content. The key is requiring both visible ‘display labels’ that users can easily see, and invisible marks like ‘digital watermarks’ embedded in metadata.
This regulation was jointly prepared by the Cyberspace Administration of China (CAC), the Ministry of Industry and Information Technology, the Ministry of Public Security, and the State Administration of Radio, Film, and Television. Authorities emphasized the need for regulation, stating that AI-generated content can lead to societal risks such as misinformation, copyright infringement, and online fraud.
Social Media Companies Respond Urgently
Major social media companies in China have quickly aligned themselves with the regulations. Tencent’s WeChat (known in Chinese as Weixin) now requires creators to specify whether content is AI-generated when uploading. If not labeled, a warning message appears for users saying “individual judgment is required.”
Douyin (the Chinese version of TikTok), operated by ByteDance, has also recommended visible labeling on all posts and added a feature to automatically record the source in metadata during the upload process.
Weibo has introduced a function to report ‘unlabeled AI content’, while Xiaohongshu (RedNote) announced it would attach labels autonomously if not self-reported.
With WeChat having over 1.4 billion monthly active users (MAU) and Douyin with 760 million users, the implementation of these regulations is expected to have an immediate impact across China’s online ecosystem.
Targeting ‘Deepfake Crimes’
Chinese regulatory authorities identified deepfake technology as a major risk. Deepfakes, which manipulate a person’s face, voice, and actions using AI, can infringe on personal honor, cause social confusion, and even pose a national security threat.
The CAC has set the enhancement of AI content management and enforcement of mandatory labeling as key tasks in the 2025 Cleansing Campaign, a nationwide initiative to rectify online order. This year’s campaign also focuses on cracking down on ‘fake marketing’ and ‘youth protection.’
A New Burden for Content Creators and Companies
The new regulations demand substantial changes from content creators. Influencers, e-commerce sellers, and marketing companies must label AI-generated texts or videos when posted. Failing to comply can lead to content deletion or penalties.
Industries frequently utilizing AI tools, such as image editing and synthesized video production in entertainment and shopping sectors, are directly affected. The regulation’s reach extends to content creating ‘virtual environments,’ including games and the metaverse, necessitating new workflows across industries.
Experts also point out technical challenges, as the automatic differentiation between AI-generated and human-created content is not yet perfect. Mistakes can occur wherein human creations are wrongly identified as AI-generated or vice versa. Consequently, companies are investing in developing real-time detection algorithms at the upload stage.
A Model Stricter Than International Regulations
This regulation is considered stricter than the European Union’s ‘AI Act’. While the EU focuses mainly on ensuring users recognize content as AI-generated, China has further mandated metadata watermarks to prevent users from arbitrarily deleting or altering the labels.
A report by the international law firm White & Case suggests that China’s regulation prioritizes national security and social order beyond simple industry management. This has raised concerns about freedom of expression and criticisms of excessive regulation.
Global Impact and Future Outlook
The measure is expected to have significant global repercussions, as it applies to the market with the world’s largest online population. Major foreign tech companies seeking to enter or partner with the Chinese market must comply with these regulations.
As AI technology advances, platforms are expected to adopt more sophisticated watermarking technologies. Within China, AI giants like Baidu and Alibaba are likely to collaborate to establish technological standards.
Ultimately, the success of the regulation hinges on the participation of users and creators. Whether they will adhere to the label mandate or attempt to evade it will be crucial to the regulation’s effectiveness.