China issues first national policy framework for AI agents

China’s Cyberspace Administration, National Development and Reform Commission, and Ministry of Industry and Information Technology jointly issued Implementation Opinions treating AI agents — systems capable of autonomous perception, memory, decision-making and execution — as a governance category distinct from general generative-AI tools, in what outside analysts described as the first national policy framework specifically written for agentic AI.

The framework requires developers to define clear decision boundaries across three tiers of autonomy: decisions reserved to the user, decisions the agent may take with prior user authorisation, and decisions the agent may take on its own. It guarantees users a right to know about, and a final say over, decisions an agent makes autonomously. High-risk applications — healthcare, transportation, media and public safety among them — face filing requirements, mandatory testing and direct regulatory oversight, while lower-risk consumer applications are left largely to platform self-governance backed by credit-scoring systems for compliance. The Opinions also call for a national standards system covering agent interfaces, data exchange, safety assurance and trustworthiness certification, aimed at interoperability across the industry, and flag anthropomorphism-driven dependence among minors and elderly users, plus automated fraud and attack tooling, as harms the framework is meant to address.

Analysts noted the framework arrived ahead of comparable agent-specific rules from Western regulators, who had so far addressed autonomous AI systems mainly through general-purpose AI legislation rather than a dedicated instrument.