China Tightens Rules on AI-Edited Videos in New Media Control Campaign

The measures suggest that online creators who use generative AI tools to remake film and television material could face administrative penalties.

A response in Chinese by ChatGPT, an AI chatbot developed by OpenAI, is seen on its website in this illustration picture taken Feb. 9, 2023. Florence Lo/Illustration/Reuters

China’s media regulator launched a month-long campaign that started on Jan. 1 to crack down on what the communist regime describes as artificial intelligence (AI)-altered video content, further tightening controls on short-form videos and AI-generated media.

On Dec. 31, 2025, Chinese state media reported that the campaign, announced by the National Radio and Television Administration, targets the use of AI to substantially modify existing images, videos, audio, or text—practices commonly known in China as “AI magic editing.”

The announcement states that the effort aims to curb content that deviates from original works, particularly those involving China’s historical themes, revolutionary Communist narratives, and depictions of regime-recognized heroes.

The notice states that platforms would be required to strengthen oversight of AI-edited content, including enforcing a “review before publication” system and addressing accounts deemed to have “prominent irregularities.”

The measures suggest that online creators who use generative AI tools to remake film and television material could face administrative penalties.

Historical, Political Focus

According to the notice, regulators are prioritizing the removal of AI-altered clips derived from the Chinese regime’s historical dramas, revolutionary films, and content portraying “heroes.”

It states that some online accounts have used AI face-swapping, synthetic voice tools, and recombined footage to transform such works into short videos that circulate widely on social media platforms.

The notice also alleges that some re-edited revolutionary film clips were being reshaped with internet-style narration and redistributed in ways that authorities claim “deviate from the original spirit” of the works. Therefore, platforms were instructed to enhance content moderation and apply classified management systems to prevent the spread of unauthorized alterations.

Growing Scrutiny

The Epoch Times spoke to several content creators based in China who only revealed their surnames for safety reasons.

Zhou, who has worked for years in short-form educational videos, told The Epoch Times on Jan. 1 that AI tools are increasingly easy to use and widely accessible, including to children.

“Many of the tools are free,” he said. “Kids use a free app that can change a face in a few taps and generate a clip in under a minute. [They] download the apps, make something, and post it right away.”

Zhou said the regulatory campaign appears aimed not only at minors but also at broader issues of content direction. AI-generated material does not always align with the regime’s officially promoted narratives, he said.

A tech worker with the surname Li, who has participated in platform content review, told The Epoch Times that the most difficult material to assess is not visual manipulation itself but changes in the underlying meaning.

“Current image recognition systems can spot [face swaps],” he said. “What’s hard to review is when the logic or message is overturned. That often still requires human judgment.”

Li noted that the Cyberspace Administration of China has issued ongoing compliance reminders to platforms over the past six months, indicating that the current campaign is unlikely to be a one-time action.

Several other creators interviewed by The Epoch Times expressed concern that enforcement boundaries could continue to expand.

Broader Trend

The Chinese regime’s restrictions on AI-generated political imagery have increased since 2024, including the removal of videos that insert the faces of political figures into entertainment content after such clips gained traction online.

China introduced formal regulations on generative AI in 2023. The new campaign marks another significant step in the regime’s content moderation, as Chinese authorities increasingly view AI-generated content as a regulatory risk.

As of publication, the Chinese regime has not disclosed how it will assess the effectiveness of the campaign or whether additional measures will follow. China observers are watching closely to see whether enforcement will expand to wider forms of derivative content and private expression.

Shen Yue contributed to this report.