China Issues Draft Regulations on AI Copyright Infringement

New rules target AI misuse of likenesses

China has released draft regulations aimed at curbing AI-driven copyright infringement, a move prompted by growing complaints that actors, social-media influencers and ordinary citizens are having their likenesses stolen and repurposed in AI “short dramas.” The guidance marks a high-profile regulatory response to a wave of deepfake and generative-video content that has alarmed performers and platform users across the country. It has been reported that the draft is intended to clarify when the use of a person’s image or performance in AI-generated content crosses into illegal copying.

What the draft reportedly proposes

Details in state and industry reporting say the draft would expand existing copyright frameworks to cover AI-generated works, assign clearer liability to content generators and hosting platforms, and impose takedown and compensation mechanisms for rights holders. Reportedly, the measures also contemplate rules on consent and attribution for likenesses used as training data or as direct inputs to generative models. Observers note that many provisions remain broad and will likely be narrowed during the consultation process.

Domestic and international context

China’s move fits a broader tightening of tech governance at home: platforms such as Baidu (百度), ByteDance (字节跳动) and Alibaba (阿里巴巴) must balance fast-moving AI deployment with new legal constraints. How will startups respond? Expect companies to accelerate compliance tools and content controls. The measures also intersect with global debates over AI training data, copyright and platform liability; while unrelated to sanctions and trade policy directly, Beijing’s rules come as Western regulators and courts wrestle with similar questions around generative models and creative rights. Observers expect a public consultation and further refinement before any final regulations take effect.