China releases world's first standard system for AI agent interoperability, issues over 2,000 identity codes

China has released seven national guidance standards, GB/Z185-2026, in Beijing, marking the world's first systematic standard system for AI agent interoperability. It aims to solve problems like incompatibility between different agents, lack of trusted identity, and barriers to cross-domain collaboration, often called 'intelligent islands.'

The standards cover seven key areas: overall architecture, agent identity codes, identity management, capability description, cross-domain discovery, collaborative interaction, and tool calling. The agent identity code is a key part, and over 2,000 identity codes have been issued to agents in important industries, giving each a unique, traceable 'digital ID' to build a trusted foundation.

Companies like Volcano Engine, Xiaomi, Kuaishou, and Lenovo helped create the standards. An English version was released at the 2026 World AI Conference, with Russian and Arabic versions to follow, aiming for global adoption.