Rischi algoritmici. Alcune questioni sulla digitalizzazione della giustizia

The recent European regulation on artificial intelligence, known as the AI Act, categorizes AI systems intended for use by judicial authorities in the investigation and interpretation of facts and law, as well as in law enforcement, as high-risk systems. However, starting from this framework, some issues arise even earlier regarding the assessment of the risks associated with the judicial use of algorithmic logic. In fact, before evaluating risks, they must first be identified—an activity that is neither simple nor automatic. In this context, numerous contradictions and challenges affect the digitalization of justice. This essay aims to illustrate the main issues, starting from the fundamentally rhetorical and dialectical nature of the judicial process.