China: AI Agent Identity Policy
Summary: China reportedly introduced a national standard for AI-agent interconnection that includes unified identity management and identity-code provisions for AI agents (2026-06-26).
Sources: 2026-09-04-ai-agent-identity-news.md, 2026-10-16-ai-agent-identity-news.md, 2026-10-18-ai-agent-identity-news.md, 2026-11-04-ai-agent-identity-news.md, 2026-11-05-ai-agent-identity-news.md, 2026-11-06-ai-agent-identity-news.md
Last updated: 2026-11-09
China reportedly introduced a national standard for AI-agent interconnection that includes unified identity management and identity-code provisions for AI agents (source: 2026-09-04-ai-agent-identity-news.md, citing South China Morning Post, published 2026-06-26). This reporting is secondary, citing Chinese official and state-media sources (source: 2026-09-04-ai-agent-identity-news.md).
This is described in the source as a country-level development closely comparable to Estonia’s proposed AI-agent personal identification codes (source: 2026-09-04-ai-agent-identity-news.md).
Needs verification
The specific content of the standard (scope, enforcement mechanism, which agencies administer it) is not detailed in the current source and needs verification against the original standard documentation.
2026-10-16 / 2026-10-18: Hong Kong SAR — HKMA agent-accountability and registry work
On 15 October 2026 (14:30 HKT) the Hong Kong Monetary Authority announced its Accountability, Resilience and Trust (ART) framework for responsible AI adoption, new AI supervisory guidance and further agentic-AI sandbox testing (GenA.I. Sandbox++). An upcoming discussion paper will explore accountability anchored in intent, a registry of material AI agents and effective intervention mechanisms. These are proposals to explore; no registry or agent-specific rule has been imposed (source: 2026-10-16-ai-agent-identity-news.md and 2026-10-18-ai-agent-identity-news.md, citing HKMA — ART Framework and HK Government release, primary). Hong Kong is a separate regulatory jurisdiction from the mainland national standard described above; the mainland and HKMA approaches are not shown to be coordinated (needs verification). The report variously writes “Resilience” (2026-10-16-ai-agent-identity-news.md) and “Resiliency” (2026-10-18-ai-agent-identity-news.md).
2026-11-04: Hong Kong SAR — HKMA Project SAGE
On 2 November 2026 HKMA launched Project SAGE and an Agentic Finance Expert Group. Work areas: practical financial use cases; verification of agent identity and delegated authority (including Know Your Agent and authentication); interoperability standards; legal questions on governance, liability and consumer protection. A companion announcement confirms publication of a discussion paper on banks’ use of agentic AI (source: 2026-11-04-ai-agent-identity-news.md, citing HKMA Project SAGE and HKMA FinTech Week announcement, primary).
Relation to earlier claim: the 15 October ART announcement (above) said a discussion paper was upcoming; it is now reported published. That is a progression, not a contradiction; the earlier proposals-only status of a registry stands. Still no adopted KYA standard, registry or binding identity requirement, and the discussion paper’s proposals were not inspected (source: 2026-11-04-ai-agent-identity-news.md). Publication hour unavailable.
2026-11-04: CAC intelligent-internet cooperation initiative
At the Wuzhen summit CAC and partners proposed international cooperation including open interoperability standards (interfaces, protocols, semantics, workflows) and a trust framework for agent discovery, negotiation and execution so cross-platform interactions can be verified and traced, plus safety provisions on data poisoning and loss of control. A cooperation proposal, not an adopted specification, operational identity registry or enforcement mechanism; no personal-identification-code requirement. Chinese-language primary text summarized in English (source: 2026-11-05-ai-agent-identity-news.md, citing CAC; Jiangsu republication is the same text). Note this does not confirm or modify the June 2026 interconnection-standard identity-code provisions recorded above.
2026-11-05: CAICT and Zetrix Know-Your-Agent cross-border pilot
Zetrix (Malaysia-linked issuer) announced with CAICT a KYA service tying verifiable agent identities to authenticated individuals or organizations, with scoped, revocable permissions, limits on onward delegation and signed activity records; a planned pilot across China, ASEAN, the Middle East and Africa keeps local trust decisions; participants unconfirmed. Findings intended for ITU-T SG17’s Focus Group on Trust and Identity for Humans and Agentic AI. Issuer claims, not independently validated; not an adopted standard or mandate (source: 2026-11-06-ai-agent-identity-news.md, citing Zetrix). See also malaysia-ai-agent-policy.