Regulatory  ·  2026-07-17

China's Implementation Opinions on Intelligent Agents take effect — first dedicated AI agent regulatory framework

RegulatoryHigh impactChina
China's Cyberspace Administration (CAC), National Development and Reform Commission (NDRC), and Ministry of Industry and Information Technology (MIIT) jointly issued the 'Implementation Opinions on the Standardized Application and Innovative Development of Intelligent Agents' on May 8, 2026; the framework entered into force on July 15, 2026. It is the world's first dedicated regulatory category for autonomous AI agents (distinct from generative AI/chatbot rules), defining agents as systems with autonomous perception, memory, decision-making, interaction, and execution capabilities. The framework applies across nineteen sectors (healthcare, transportation, finance, manufacturing, education, government services, public safety, etc.), establishes a three-tier decision-authorization structure classifying agent actions by consequence level with scaled human-approval thresholds, requires filing/registration and compliance testing for agents deployed in high-risk/sensitive sectors, and empowers sector regulators to apply measures including product recall for non-compliant agents.
This is the first jurisdiction-specific binding framework anywhere treating autonomous AI agents as a distinct governance category with concrete operational obligations (tiered authorization, mandatory filing, recall powers, kill-switch-style traceability). Any company deploying agentic AI systems that touch Chinese operations, users, or the 19 covered sectors is now subject to a live compliance regime, not a discussion draft. It sets a global precedent that other regulators (evidenced by parallel activity from Illinois, Singapore's IMDA, and the EU) are likely to reference or react to.
Companies with AI agent deployments touching Chinese markets/operations must verify compliance with the three-tier decision-authorization structure, complete required filings for agents operating in the 19 designated sectors (especially healthcare, finance, transportation, public safety), and establish traceability/version-control and recall mechanisms as directed by sector regulators.
Digital Policy Alert — Global Digital Policy RoundupTechTimes — Ant Group Open-Sources Agent Security ToolForbes — China Plans AI Agent Recalls
See this in the live feed Explore related AI security and governance findings — updated every morning.
Open the feed →