Abstract

The widespread use of AI in human activities, both public and private, is desirable, as long as the fundamental rights and freedoms of human beings are not sacrificed. Administrative jurisprudence has identified three rules: 1) the full knowability/understandability of the form and criteria used by the AI; 2) the principle of non-exclusivity of the algorithmic decision; 3) the principle of non-discrimination. The structured use of algorithms brings with it the need to identify who is responsible for any damage: it is a joint liability, attributable to different degrees of fault, between the creator (producer) of the algorithm, the computer technician (for use of potentially dangerous tools), the owner of the database (custodian of the data provided to build the algorithm) and the official who benefits from the use (custodian and user).