From algorithmic legality to artificial (administrative) legality. Preliminary reflections for a study

Abstract

The advent of artificial intelligence, now in the form of generative AI, such as ChatGPT, poses formidable problems and at the same time offers extraordinary opportunities for improving public action. The question, from the point of view of administrative law, is to understand the phenomenon and therefore reflect on which rules can continue to guarantee the individual in the relationship with the new digital power. Doctrine, legislator and, above all, jurisprudence, in recent years have outlined the foundations of a new algorithmic legality: the essay reflects, in introductory terms, on the “hold” of those principles (of transparency, non-exclusivity, non-discrimination) and on their actual “consistency” in the face of the new prospects of “artificial administration” (meaning by this the administration that uses, for its decisions, advanced artificial intelligence solutions).