China releases national standards for AI agent interconnection, covering 7 core specifications

China's State Administration for Market Regulation held a press conference in Beijing to announce the official release of a national standard series for AI agent interconnection. The series includes 7 parts, creating a unified framework for the entire process of agent connectivity, aiming to solve issues like inconsistent communication interfaces, lack of identity management, and messy collaboration rules. China's AI agent industry is growing fast, with successful cases in smart manufacturing, smart government, and smart transportation, but cross-platform needs are rising, so standardization is urgently needed to break down barriers.

The 7 standards cover core areas like overall architecture, identity codes, identity management, agent description, agent discovery, agent interaction, and tool calling, forming a complete chain from identity to tool use. The Ministry of Industry and Information Technology revealed that over 100 experts from 70 leading companies joined the drafting, with 600 public comments received. A prototype system was built to verify core functions. Over 100 companies have joined a joint initiative, and more than 50 are running pilot projects.

Beijing's Haidian district is taking the lead in implementing the standards, with local companies like Volcengine, Xiaomi, Kuaishou, and Lenovo deeply involved. Next steps include pilot projects in integrated circuit design, embodied intelligence, and industrial software, using real scenarios like government services and urban management. The regulator also noted that other AI standards in development cover frontier tech like embodied intelligence and world models, as well as infrastructure and industry-specific applications, to build a complete national AI standards system.