Gaps in criminal liability arising from machines equipped with artificial intelligence

This paper examines the gaps in criminal responsibility arising from situations involving machines that have been endowed with artificial intelligence (AI). Taking into consideration the requirements stemming from the principle of criminal legality, the fragmentary nature of criminal law, and the imperatives of human dignity, this study adopts a restrictive approach to the concept of a “responsibility gap,” setting it apart from other instances of non-liability. Starting from the premise that intelligent artificial entities are not subject to criminal liability by legislative decision, it is argued that such a gap emerges when the perpetration of a legally relevant act results from the interaction between two agents: one that can indeed be held criminally accountable (such as a natural person or a corporate entity) and an intelligent artificial entity, yet no liability can be imposed on the former due to a technical characteristic inherent in the intelligent machine’s technological configuration.