GIUSTIZIA PENALE E DECISIONI ALGORITMICHE PROSPETTIVE DEI RISK ASSESSMENT TOOLS NORDAMERICANI TRA PROGRESSO TECNOLOGICO, INTELLIGENZA ARTIFICIALE E GIUSTO PROCESSO

The main goal of this research is to assess the compatibility between the use of “predictive tools” in the criminal process and the protection of due process principles and values. In particular, the research primarily focuses on algorithmic risk assessment tools that assist judges in assessing the defendant’s dangerousness and likelihood of recidivism. The aim is to take into consideration the advantages of applying these algorithmic tools in some key junctures of criminal proceedings, both in terms of efficiency and objectivity and uniformity of decisions. At the same time, it is of crucial importance to carefully consider the risks associated with the use of algorithmic systems with regard to the fundamental tenets of due process and fair trial rights (and in particular, right to defense counsel, the protection against self- incrimination, the presumption of innocence, and judicial impartiality). The research will be carried out through a comparison between the US and Italian legal systems. In the United States, the use of AI, big data, and machine learning tools in legal practice has now become a reality, and in recent years we have been witnessing a noticeable growth in the use of algorithmic risk assessment tools at all stages of the criminal process. In the Old Continent, on the other hand, the use of algorithmic tools has not yet reached the level of diffusion comparable to U.S. jurisdictions, and the first official reports and policy documents on the development of algorithmic and AI systems have only recently been published. This research aims to assess whether similar tools, if undestood and corrected, could be implemented also in the Italian criminal justice, and what “natural defenses” the Italian system has for protecting the fundamental rights of individuals involved in the assessment of these tools.