WAIC Unveils a Global Initiative on Agent Interoperability: Ten Rules for Multi-Agent Collaboration
At WAIC 2026's high-level AI governance meeting, China's CAC and partners proposed a global initiative on agent interoperability, spanning ten areas including R&D, standards, safety ethics and cross-border data.
A global initiative for "agent interoperability"
Per IThome (July 17), at WAIC 2026's high-level meeting on global AI governance in Shanghai, China's Cyberspace Administration and partners formally proposed a "Global Cooperation Initiative on Agent Trust, Connectivity and Interoperability," aiming to unlock agents' potential for sustainable development, prevent an "intelligence divide," and build an open, trustworthy, safe and inclusive agent ecosystem. The initiative sets action directions across ten areas.
Technology, standards and infrastructure
- R&D: greater investment in common enabling tech such as agent identity and mutual recognition of collaboration, moving multi-agent systems from "controlled interaction" to "autonomous collaboration" with efficient, scalable, low-latency cross-platform interoperability;
- Standards: developing open, non-discriminatory, transparent international interoperability standards, aligning interfaces, protocols, semantics and process layers to avoid technical barriers and ecosystem fragmentation;
- Infrastructure: jointly building open, shared "intelligent internet" infrastructure.
Safety ethics and cross-border data
The initiative calls for a full-lifecycle safety governance framework embedding safety and ethics into R&D, deployment and operations, never crossing lines of safety or human dignity; on data flow, it respects differing national data-security and privacy regimes and advocates an open, inclusive, non-discriminatory cross-border data framework. It also encourages providing open-source tools, training and funding to developing countries, fostering "South-South and South-North" digital cooperation.
Why it matters
As agents move from "chatting tools" to economic units that autonomously find partners, collaborate and transact, their trusted identity, controllable permissions, auditable behavior and cross-platform interoperability are becoming the base-layer questions for whether an agent economy can scale. This initiative raises them to the level of global governance — alongside the concurrently released "AI Cooperation and Development Action Plan" and "International AI Ethics Governance Action Plan," an attempt to "set the rules first" for the agent era.