China sets legal liability rules for AI deepfakes and digital cloning

China’s Supreme People’s Court has issued new guidelines establishing legal liability for the misuse of artificial intelligence, specifically targeting deepfakes and digital cloning. Under these rules, creating or distributing AI-generated replicas of a person’s face or voice without their consent is considered a legal infringement, particularly when used to spread falsehoods or damage reputations.

The regulations also address the responsibilities of AI service providers. Companies may face legal consequences if they fail to act promptly after being notified that their models have generated content infringing on individual rights. This includes instances where AI models produce inaccurate or invented information, a phenomenon known as “hallucinations.”

Additionally, the court addressed algorithmic fairness in commerce. The guidelines prohibit businesses from using algorithms to implement discriminatory pricing, such as charging different customers different rates for the same product or service without justification. Zhou Jiahai, director of the Supreme Court’s investigation office, emphasized that the law must act quickly to protect consumer interests because not every consumer can be an expert in detecting digital deception.