Governing Artificial Intelligence: The RICE model

Abstract

The integration of AI in the exercise of public authority and decision-making processes that threaten fundamental rights, human health and safety, as well as democracy, the rule of law, and the environment requires not only rapid and statistically accurate measurements, but also governance tools and practices to illuminate internal processes and the corresponding logic, predict situations where the probability of failure increases, and ensure the full effectiveness of the principle of accountability.   In this phase, where a paradigm shift involving both epistemological and operational dimensions is necessary, the Fondazione Ente Nazionale per l'Intelligenza Artificiale (ENIA) proposes and further develops the RICE Model as a solid and integrated framework suitable for supporting the compliance of AI systems with the standards required by European legislation, facilitating the incorporation of social and environmental concerns, and fostering greater trust among those directly affected by the operation of these technologies and among the public.