What the study found
The paper presents an article-by-article and layer-by-layer reference mapping from selected high-risk obligations in the EU AI Act to cryptographic and trust-service primitives. It frames this as a way to produce independently verifiable evidence about AI system behavior.
Why the authors say this matters
The authors say the mapping is relevant because it supports evidence-based verification for high-risk AI systems. They present the work as an implementation-agnostic reference, and they note that the example system used does not itself certify legal compliance.
What the researchers tested
The working paper maps obligations in Regulation (EU) 2024/1689 to primitives drawn from eIDAS/eIDAS 2.0, ETSI EN 319-series standards, IETF RFC 3161 timestamping, W3C Verifiable Credentials, JSON canonicalization, and post-quantum signature practice. It uses the Agent Trust Framework (EATF) as a worked example because its artifacts are publicly observable.
What worked and what didn't
The abstract says the mapping was completed as a reference framework and that version v1.0-preprint adds a structured failure-case appendix for broken evidence packages, completes a URL verification pass, and tightens claim-risk wording from compliance guarantees toward evidence-support formulations. It does not report experimental performance results or comparative outcomes.
What to keep in mind
The paper is a v1.0 preprint and working paper, not a finalized standard. Tyche Institute is described as a research entity, not a trust service provider or qualified trust service provider, and the paper explicitly says it does not claim that EATF or any implementation certifies legal compliance.
Key points
- The paper maps selected high-risk EU AI Act obligations to cryptographic and trust-service primitives.
- It aims to support independently verifiable evidence about AI system behavior.
- The mapping draws on eIDAS/eIDAS 2.0, ETSI EN 319-series standards, RFC 3161 timestamping, W3C Verifiable Credentials, JSON canonicalization, and post-quantum signature practice.
- EATF is used only as a worked example because its artifacts are publicly observable.
- The abstract does not report experimental performance or validation outcomes.
Disclosure
- Research title:
- Reference mapping links EU AI Act obligations to trust-service primitives
Get the weekly research newsletter
Stay current with peer-reviewed research without reading academic papers — one filtered digest, every Friday.