Semantic web framework unites evidence for eu ai rights assessments

A Reusable Semantic Web Framework for Evidence-Grounded Fundamental Rights Impact Assessments under the EU AI Act

Computers and SocietyArtificial Intelligence

Summary

The EU AI Act requires careful checks on how AI systems might affect fundamental rights before using them, but evidence needed for these checks is spread out and hard to combine. The authors created a reusable computer framework that gathers this scattered evidence into one place, focusing on AI in jobs and public services. They prepared a detailed collection of cases, annotated them with the help of AI tools, and turned everything into a format that can be searched efficiently. Their tests showed that automatic classification tools had trouble accurately sorting this evidence, highlighting challenges in relying on AI for fairness checks. All their tools and data are shared openly to help regulators and businesses use them.

What this means in practice

  • For regulatory compliance teams: Use the framework to gather and query evidence when performing fundamental rights assessments for high-risk AI systems under the EU AI Act.
  • For small and medium enterprises: Access an open knowledge graph to support compliance with AI regulations in employment and public service domains before deploying AI tools.

Authors

Faith Olopade, Delaram Golpayegani, David Lewis

Abstract

The EU AI Act (Art. 27) requires deployers of high-risk AI systems to conduct Fundamental Rights Impact Assessments (FRIAs) before deployment, yet the evidence needed for credible assessments is fragmented across incompatible incident repositories, risk vocabularies, and legal texts. We present a reusable Semantic Web-based framework that consolidates this evidence for two high-risk public sector categories: employment and worker management (Annex III(4)) and access to essential public services (Annex III(5)(a)). A curated 150-record corpus is annotated along four axes using keyword, LLM, and hybrid methods and serialised as a SPARQL-queryable knowledge graph of 1,351 RDF triples. Five FRIA demonstration scenarios surface 103 records (68.7% coverage). Evaluation against a 69-record gold standard reveals that LLM-assisted classification of the employment domain achieves only $κ= 0.045$, a cautionary result for automated fairness-related evidence retrieval in this domain. All artefacts are released openly to support adoption by regulators, national authorities, and SMEs.
View PDFOpen arXiv