Remote Attestation for Agentic AI (RA-AAI)
fg-tida-themes-remote-attestation-for-agentic-ai-ra-aai-41
| Type | FG-TIDA document |
|---|---|
| Group | Trust and Identity for Humans and Agentic AI |
| State | Submitted |
| Authors | Not entered |
| Last updated | 2026-10-23 15:10 UTC |
Formats and resources
Abstract
Remote Attestation for Agentic AI (RA-AAI)
RA-AAI
Proposer(s)
Henk Birkholz (Fraunhofer) and Houda Labiod (Huawei)
Summary
Agentic AI is rapidly evolving from prompt-based assistants into autonomous systems that plan, call tools, execute tasks locally or remotely through inter-agent collaboration and interact with real-time environments while reducing direct human control over internal decision-making logic. Facing new and complex large-scale security and privacy vulnerabilities, Remote Attestation is one major technology that could be used to provide cryptographic proof of trustworthiness of an execution environment. Remote Attestation (RA) for AI Agents is needed to prove what code+data+policy the agent is running, not just the hardware. This WG aims at analyzing the major RA’s challenges for AI Agents, identify the existing solutions provided by different SDOs (IETF, TCG, Global Platform) and the gaps to cover the attestation and verification needs of an AI agent.
Scope / Open Questions
We aim to answering to important fundamental questions:
- What to attest?
- How to attest?
- When to attest?
Related Work
RA standards in IETF (several Working Groups including RATS, OAuth, WIMSE), TCG, Global Platform.
Related Themes
Agent ID, Trust management for Digital ID across borders, Agent identity and trust stack
Discussion
Revisions
- No file revisions have been uploaded.