The complex multilevel regulation of artificial intelligence: between subjectivity, innovation and market.

Abstract

The development of computer science, algorithms and artificial intelligence systems with reference to public and private relationships highlighted the dual aspect of the need to mediate between the demands of technological and market development and, at the same time, to provide adequate protection to people and their fundamental rights, evaluating the risks in balance with the undoubted advantages. It is necessary to ask ourselves whether, according to AI, it is appropriate to resort to common civil liability regimes or to refer to different special legislation; and also evaluate whether to invest AI with legal subjectivity and therefore the possibility that an automated system will be directly liable for the damage caused. On a dogmatic level the European legislator considers the inadequacy of national regimes for the attribution of fault-based liability: they would not, in concrete terms, be suitable for guaranteeing effective protection to subjects damaged by the use of AI systems, due of the complexity, opacity and autonomy that characterizes them.