China-led initiative on building inclusive and open AI agent ecosystem issued

A China-led initiative on mutual trust and interconnectivity among artificial intelligence agents was released on Friday, aiming to build broad consensus and work with global partners toward an open, trustworthy, secure, and inclusive agent ecosystem.

The plan was put forward by the Cyberspace Administration of China, in collaboration with relevant authorities, during the main forum of the 2026 World AI Conference and High-Level Meeting on Global AI Governance in Shanghai.

As one of the most transformative technological forms of the AI era, agents — in terms of their mutual trust, interconnectivity, and interoperability — are profoundly shaping the integration of the global digital economy and the future of human society. The initiative is designed to unlock the potential of agents in advancing sustainable development while preventing the emergence of an intelligence divide, the administration said.

The measure calls on countries to increase research and investment in foundational technologies such as agent identity authentication and collaborative mutual recognition to solidify the technological bedrock of the field. It also supports the development of open, non‑discriminatory, and transparent international standards for agent interoperability, together with an open‑source mechanism for standard validation and iterative refinement.

Under the initiative, all parties are encouraged to jointly explore interconnection and collaborative architectures for agents and work together to build open and shared intelligent internet infrastructure. At the same time, the outcomes are expected to be broadly applied to scientific research, industrial development, consumption stimulation, public well‑being, and social governance — with particular priority given to leveraging them in addressing global challenges such as climate change, public health, educational equity, and poverty alleviation.

China welcomes the broad participation of developed and developing countries, large enterprises and micro, small and medium-sized entities, as well as research institutions and open-source communities, in advancing the building of mutual trust, interconnectivity, and interoperability among agents. It also calls for embedding safety and ethical considerations into technological research and development to jointly uphold bottom-line security.

In addition, the plan underscores the importance of promoting data flows, strengthening resource sharing, and bridging the intelligence divide, with the aim of developing cross-cultural, cross-linguistic, and cross-disciplinary norms for agent collaboration.

Furthermore, helping minors, the elderly, people with disabilities, and low-skilled groups enhance their AI literacy has also been incorporated into the initiative.

Guided by the principle of extensive consultation, joint contribution, and shared benefits, the administration said that China sincerely invites all parties that subscribe to the above vision to join hands — with an open and inclusive mindset and prudent yet pragmatic actions — to jointly foster the development of the agent ecosystem and contribute wisdom and strength to the building of a community with a shared future in cyberspace.