- The initiative proposes AI-agent interoperability standards and a lifecycle trust framework.
- Its infrastructure and safety measures remain proposals rather than demonstrated deployments.
A proposal for agent cooperation
China's cyberspace regulator has proposed international standards for cooperation among AI agents as part of an intelligent-internet initiative published on November 4. The Cyberspace Administration of China says it put forward the initiative with relevant parties at the Wuzhen summit. The document contains eight areas for cooperation, rather than announcing an implemented technical standard.
It calls for open, non-discriminatory and transparent interoperability standards, with compatible interfaces, interconnection protocols, semantics and processes. A proposed trust framework would cover agent discovery, negotiation and execution, aiming to make cross-platform and cross-sector cooperation verifiable and traceable.
Infrastructure and risk controls
The initiative also advocates coordinated deployment and interconnection of infrastructure including IPv6, advanced mobile networks, satellite internet and intelligent computing. Its application proposals include embodied intelligence, humanoid robots and services built around models or agents.
On safety, the document identifies data poisoning and loss of control over agent behaviour as risks to address. It proposes incorporating security and ethics throughout research, deployment and operations, alongside information sharing and emergency cooperation. These are cooperation proposals; the text does not establish that interoperable systems or the proposed safeguards have already been delivered.
Sources & context
Go to the original material. Company claims remain attributed to their sources.
01
