China released the AI Safety Governance Framework 3.0
In September 2026, China published the AI Safety Governance Framework 3.0, which addresses risks of AI systems including AI agents and the misuse of CBRN technologies. It builds on the 2022–2023 rules for algorithms, deepfakes, and generative AI.
In September 2026, the Chinese government published the AI Safety Governance Framework 3.0 – a document that, according to the source article, describes an approach to identifying, classifying, and addressing risks of AI systems throughout their entire lifecycle. The framework builds on previous Chinese regulation: the 2022 rules for algorithmic recommendation services, the measures for synthetic (AI-generated) content effective from January 2023, and the interim measures for generative AI services from August 2023.
The new version covers risks associated with AI models and data, vulnerabilities of open source models, the impact of AI on employment and society, and increasingly autonomous systems. It includes a separate mechanism for managing risks associated with rapidly developing AI agents. According to the article, the framework also addresses topics that are currently raising concerns in the US – highly autonomous systems, misuse of chemical, biological, and nuclear technologies, and the risk of losing human control over advanced AI. It proposes ongoing safety evaluations, risk-graded measures, maintaining human oversight, and emergency intervention mechanisms.
An independent analysis by the American think tank Carnegie Endowment for International Peace describes the framework as China's attempt to "advance frontier AI safety without the need for hard regulatory interventions." The article contrasts the framework with calls from American AI leaders, such as Dario Amodei of Anthropic, for new regulatory mechanisms – while, according to the author, similar mechanisms (evaluations, corporate accountability, risk classification, algorithm registration, technical standards, state oversight) have already been applied in China for years.
Why it matters
The framework shows that alongside the AI regulation debate led by American companies, there already exists an established, gradually built Chinese regulatory system that also covers the latest risks, such as autonomous AI agents. For companies operating in the Chinese market or integrating Chinese AI models and services, this means a concrete set of requirements (security evaluations, algorithm registration, agent risk management) that must be taken into account in compliance and in managing the AI technology supply chain.
Relevant practical impact
What this means
For a business
Companies developing or deploying AI products with ties to the Chinese market or Chinese suppliers must deal with an existing regulatory framework (security evaluations, algorithm registration, specific oversight of AI agents), which is a different and in some respects more elaborate system than the regulation currently being discussed in the US.
Risks and complianceCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.