China’s rules governing how AI agents make decisions took effect nationwide on July 15, following the “Implementation Opinions on the Standardized Application and Innovative Development of Intelligent Agents,” which the Cyberspace Administration of China (CAC) issued jointly with the National Development and Reform Commission and the Ministry of Industry and Information Technology on May 8.
What the rules require
The document defines an AI agent as a system with autonomous perception, memory, decision-making, interaction, and execution abilities — distinguishing it from a standard chatbot that only responds to prompts. It sorts agent actions into three tiers: decisions users must make themselves, decisions an agent may take only with explicit user authorization, and a narrow set of low-risk actions an agent can carry out on its own. Regulators say users must retain final decision authority over consequential actions.
Developers must also embed rule-based behavioral guardrails limiting what agents can do in the physical and digital world, and keep verifiable, traceable logs of agent behavior in significant applications, so what an agent did — and why — can be reconstructed afterward.
Filing required in sensitive sectors
Organizations deploying agents in healthcare, transportation, media, and public safety must complete a formal filing with regulators, undergo compliance testing, and follow defective-product recall procedures before wide release. Analysts tracking Chinese tech law describe the framework as the first dedicated, nationwide regulatory category built specifically around autonomous AI agents, rather than generative AI output in general.
Part of a broader push
The rules sit inside China’s “AI+” initiative, a State Council-backed plan that sets a target of over 70% penetration for intelligent agents and next-generation smart terminals by 2027. A separate rule restricting AI companion and emotional-chatbot services also took effect the same day, though it covers a different category of product.