China wants a single AI law, but has no deadline yet

China's State Council has set a goal in its 2026 legislative work plan of speeding up comprehensive legislation for the healthy development of artificial intelligence and improving AI governance. This is according to a reproduction of the plan's introduction.

The plan also mentions legislation on the shared building blocks of AI: data, computing power, algorithms, intellectual property rights, network security and supply chain security. In addition, there are rules for key areas of application. The reproduced passage does not set a date for when the law should be completed.

Proposals and assessments

In June 2025, delegates to the National People's Congress, among them Quan Heng, proposed a comprehensive AI law. The proposal included risk categories for prohibited, high-risk, transparency-risk and low-risk systems, impact assessments and liability rules.

In May 2025, Professor Zhang Linghan of the China University of Political Science and Law pointed out that the 2025 legislative plan had changed an AI law from a "preparatory item" to "advancing legislation." According to her, the change reflects strategic considerations and not less urgency.

Rules already exist

China is not without rules today. Since 2022, Chinese tech companies have been required to give the authorities information about their algorithms and keep it updated as the technology evolves. The rules for generative AI from July 2023 require that output that is publicly available comply with the state's information controls. According to the New York Times, the first drafts were softened after lobbying from the industry.

In December, a draft of rules for AI companion chatbots was presented. They are meant to prevent dependency and require companies to build emotional user profiles that can be used to intervene. In June 2025, Shanghai's cyberspace authority also announced enforcement actions against AI providers that violated the rules on generative AI and personal data.

Zhipu AI warned investors in its IPO filing that the rules could have significant legal, financial and operational consequences.

According to Scott Singer of the Carnegie Endowment, China's rules focus more on information control and data protection than on safety scenarios, as in California's AI law. Graham Webster of Stanford assesses that Chinese regulators will govern AI, but weigh enforcement against whether it could hamper entrepreneurship.

Travel restrictions for key personnel

Bloomberg Law reports that China is restricting overseas travel for key personnel in advanced AI at private companies such as Alibaba and DeepSeek. They must obtain approval from the authorities before traveling abroad. Bloomberg describes it as an extension of a long-standing practice for key employees. The travel restrictions are not directly part of the AI law.