|
Work item:
|
X.AIA-dii-req
|
|
Subject/title:
|
Security requirements for DLT-based decentralized identity interoperability across artificial intelligence agents
|
|
Status:
|
Under study
|
|
Approval process:
|
TAP
|
|
Type of work item:
|
Recommendation
|
|
Version:
|
New
|
|
Equivalent number:
|
-
|
|
Timing:
|
2028-12 (Medium priority)
|
|
Liaison:
|
ITU-T SG13, SG20, SG21, IETF, IEEE, ISO/IEC JTC1/SC27 , ISO/TC 307
|
|
Supporting members:
|
State Grid Corporation of China; China Unicom; ZTE Corporation
|
|
Summary:
|
Artificial intelligence (AI) agents are autonomous, networked software entities that act on behalf of users, services, or organizations. Their digital identity—comprising attributes, credentials, attestations, and provenance bindings—defines who an agent is, what it is authorized to do, and which human or system it represents. Identity interoperability requires standardized data models, minimal attribute sets, and supporting infrastructure—such as federation, trust anchors, credential exchange protocols, onboarding services, and lifecycle management—that enable agents from different vendors, domains, and jurisdictions to authenticate, negotiate trust, and enforce policy without bespoke integration. Interoperability should also support end‑to‑end identification, authentication, and authorization across the interaction path to ensure accountability and security.
Security requirements for decentralized identity interoperability (DII) across AI agents should therefore treat identities and supporting infrastructure as a unified problem. Requirements need to guarantee verifiable, tamper‑resistant identity assertions and bindings; secure provisioning, key and credential management; consistent revocation and recovery semantics; interoperable trust and federation behaviors; and end‑to‑end authorization that enforces least privilege and safe delegation. They should also mandate auditability and provenance for accountability, privacy‑preserving disclosure controls, and a threat/assurance framework to address impersonation, supply‑chain compromise, and cascading failures across domains.
This draft Recommendation identifies introduces an end-to-end decentralized identity interoperability across AI agents with heterogeneous identity solutions, identifies relevant security risks and threats, and specifies the security-enhanced framework and requirements.
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2026-06-10 10:17:06
|
|
Last update:
2026-06-10 10:20:11
|
|
