Let's have a chat with the EU AI Act
Adam Kovari, Yasin Ghafourian, Csaba Hegedus, Belal Abu Naim, Kitti Mezei, Pal Varga, Markus Tauber
TL;DR
The paper addresses the challenge of navigating complex AI regulations by proposing an AI-driven self-assessment chatbot powered by Retrieval-Augmented Generation to retrieve up-to-date regulatory texts from both public and proprietary sources. It systematically compares Naive RAG and Graph RAG architectures, detailing indexing, retrieval, and generation workflows, as well as evaluation strategies including expert-grounded validation and industry use cases. The system processes multimodal regulatory documents through OCR and embeddings to deliver real-time, context-aware compliance guidance with source attribution, aiming to streamline regulatory adherence and governance. Future work envisions broader regulatory coverage, the integration of legal checklists, and deeper Graph RAG capabilities to handle more complex prompts and scenarios in responsible AI development.
Abstract
As artificial intelligence (AI) regulations evolve and the regulatory landscape develops and becomes more complex, ensuring compliance with ethical guidelines and legal frameworks remains a challenge for AI developers. This paper introduces an AI-driven self-assessment chatbot designed to assist users in navigating the European Union AI Act and related standards. Leveraging a Retrieval-Augmented Generation (RAG) framework, the chatbot enables real-time, context-aware compliance verification by retrieving relevant regulatory texts and providing tailored guidance. By integrating both public and proprietary standards, it streamlines regulatory adherence, reduces complexity, and fosters responsible AI development. The paper explores the chatbot's architecture, comparing naive and graph-based RAG models, and discusses its potential impact on AI governance.