This project addresses the challenge companies face in complying with ethical principles and legal rules for Artificial Intelligence (AI) systems. Currently, there is a lack of technical tools to verify the compliance of such AI systems according to ethical and digital trust criteria.

This project is a preliminary feasibility study for a future project that aims to develop an AI program that can be integrated into corporate decision-making tools. This program would allow companies to self-check the compliance of their AI projects using personal data with cross-cutting criteria (digital trust and ethics). The check would be performed by an integrated program (using machine learning) which would be enriched by the experience of compliance checks and several external resources.

This study starts with a review of the AI regulatory framework and a clear identification of compliance needs with partner companies before proposing a technical solution. It then examines existing methods for verifying compliance and finally proposes a model for considering ethical AI systems.