Revision 2.4 · State-endorsed identity concept
A proposed trust layer for agents that exchange information across institutions: an issuer attests an identity or role, an operator binds that evidence to an agent, and policy controls every permitted interaction.
This page maps enabling legal and policy foundations—not a common international A2A system. The sources below support digital identity, standards, safety, and interoperability in different ways. Each jurisdiction retains its own law, sovereignty, authorization rules, and data-protection requirements.
Identity should travel with verifiable authority, bounded purpose, and an audit record. It should not turn an AI model into a legal person, grant a general right to access information, or bypass a human or agency decision.
Issues a signed identity, organizational role, or credential under applicable law and assurance requirements.
Registers a specific agent, its model version, endpoint, owner, permitted tools, and revocation route.
Evaluates jurisdiction, consent, scope, assurance, data classification, and time before release.
Accepts only the minimum authorized claim or result and records provenance for review and redress.
Policy starting points
These sources are design inputs for the C=US concept. They are not evidence that the jurisdictions have adopted a shared registry or mutually recognized agent credential.
Binding digital-identity framework
Regulation (EU) 2024/1183 establishes the European Digital Identity Framework. It provides for interoperable wallet functions, common protocols and interfaces, and secure sharing of person-identification data and electronic attestations.
Conceptual use: an EU-compatible issuer/verifier pattern for attributed roles and selective disclosure. The regulation concerns digital identity wallets; it is not an A2A authorization mandate.
Federal policy and voluntary standards
The 2025 AI Action Plan promotes AI standards and adoption, while the NIST AI Risk Management Framework provides voluntary governance and risk-management guidance.
Conceptual use: a standards-and-assurance baseline for agent operators. No national U.S. statute establishing a general state-endorsed A2A identity layer is asserted here.
Federal digital-service policy
Canada’s Guideline on Service and Digital directs departments to use approved trust frameworks and support interoperable identity assurance; its enterprise architecture calls for secure APIs and appropriate authentication of individuals, processes, or devices.
Conceptual use: federated trust and secure service interfaces. Canada’s Auditor General has also reported that a national interoperable digital-identity approach remains incomplete.
AI safety and governance policy
China’s AI Safety Governance Framework and its Global AI Governance Action Plan address safety governance, traceability, information sharing, and international cooperation.
Conceptual use: traceability, risk evaluation, and governed information exchange. These materials do not create a cross-border A2A identity-recognition mandate.
Proposed A2A information flow
Design safeguards
An identified organization remains responsible for each deployed agent and high-impact decision.
Use selective disclosure, purpose limits, retention rules, and data-classification controls.
Local law and the receiving system’s policy govern every release; no credential overrides them.
Credentials, endpoints, and permissions must be rapidly revocable and independently reviewable.
Primary sources