Sui rapporti tra pena, prevenzione del reato e prova nell’era dei modelli computazionali psico-criminologici

Abstract

The paper is based on some recent effects of the digital turn upon a rather classic branch of research, both theoretical and empirical, such as individuals’ dangerousness. It is a multifaceted research area, between criminal law, jurisprudence and criminology, in which one of the goals has always been creating quantitative models to predict risk (of recidivism and violent behaviour): today, such models can benefit of the
most recent achievements in computer science. However, it is crucial to reconstruct the basic lines upon which the phenomenon developed over the decades, in order to avoid the ambiguous approach (fascination and dystopia, at the same time) that currently surrounds computer science and, in particular, AI.