ARTIFICIAL INTELLIGENCE AND CRIMINAL LAW. CHALLENGES TO SOME TRADITIONAL CATEGORIES

The challenges that technological evolution has always posed to legal formants now find a new object in the emergence of artificial intelligence. The concept embraces a multiplicity of systems, operating in the digital sphere or even in the physical world, if equipped with hardware, such as robots or self-driving vehicles, in any case interacting with human beings and the environment. Together with immeasurable benefits for individuals and the community, however, they also create new risks, due to the decision-making autonomy that characterises them to varying degrees, based on autonomous learning mechanisms from the web and the external environment. Hence a diaphragm is interposed between the act of man 'behind' such machines and their behavior or effects, which may be 'unpredictable' and offend legal goods, including fundamental rights, deserving of criminal protection. Thus the need to adapt the criteria for attributing liability to the natural and legal persons in whose interest they operate, while at the same time respecting the guarantee principles of criminal law, in particular of legality and of culpability. In this regard, the recommendations on substantive criminal law that the Association Internationale de Droit Pénal has drawn up for the forthcoming international congress in Paris in 2024, dedicated to the issue of criminal justice in the face of artificial intelligence, are taken into account