Artificial Intelligence and Criminal Compliance: Toward A Unified Model of Risk Governance

In contemporary risk society - characterized, as the most sophisticated sociological and legal theory has shown, by a structural transformation of collective dangers and by the resulting centrality of preventive mechanisms - criminal law is compelled to rethink its foundational categories in light of a paradigm no longer grounded in the harmful event, but in the capacity to govern complexity ex ante. Within this scenario, the original prevention oriented system (i.e. risk society), which has progressively shifted the axis of responsibility from psychic interiority to the quality of organizational structures, provides the ideal context in which to explore the possibility of constructing a model of criminal compliance capable of adapting to any potential author of offence: the natural person, the collective entity, and, today, the algorithm