EUROCOMPLY: Enabling zero-touch AI compliance auditing via LLM-based agentic AI

Ameur, Mazene; Brik, Bouziane; Ksentini, Adlen
IEEE Communications Magazine, 2025

In this paper, we present EUROCOMPLY, a novel framework designed to automate regulatory compliance verification in Artificial Intelligence and Machine Learning (AI/ML) systems for the telecommunications sector. With the increasing adoption of AI/ML, ensuring adherence to the European AI Act (EU AI Act) and the General Data Protection Regulation (GDPR) has become critical to avoid deployment delays and legal penalties. EUROCOMPLY leverages Agentic AI to inspect datasets and AI/ML pipelines for alignment with the EU AI Act, the GDPR, and 3GPP AI/ML-related standards. The framework employs a dual-mode retrieval architecture that combines vector-based and graph-based retrieval for enhanced regulatory interpretation. We validate EUROCOMPLY on 20 telecom use cases across four realistic datasets, demonstrating high faithfulness and strong performance through expert assessments and LLM-as-a-Judge evaluations.


Type:
Journal
Date:
2025-11-18
Department:
Communication systems
Eurecom Ref:
8507
Copyright:
© 2025 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

PERMALINK : https://www.eurecom.fr/publication/8507