AI Systems Involved in Harmful Events: Liable Persons or Mere Instruments? An Interdisciplinary and Comparative Analysis

Abstract

The article investigates the nature of AIs under criminal law, i.e., whether they are legal persons or mere tools. The study applies a double methodology. Firstly, from a comparative perspective, it analyses the US and the Italian legal systems, as they represent the two main legal traditions in the Western World, namely, common law, and civil law. Secondly, it applies the interdisciplinary research method, by reference to non-legal disciplines. The article criticizes the doctrine maintaining that AIs may be considered as legal persons. Then, it aims to demonstrate the opposite thesis, according to which AIs are mere tools.