AI and bad robots

摘要

This chapter aims to cast light on the normative side of the law. It focuses on how the discipline of bad robots actually works in the field of criminal law. The chapter illustrates today's state-of-the-art in criminal law, and why the level of robotic autonomy is at times sufficient to produce relevant effects in the field of, say, contractual obligations, but arguably is insufficient to bring today's robots before judges and have them declared guilty in criminal courts. It explores the scenario of new robotic crimes, by considering both hypotheses of robots with a criminal mind of their own, and new types of robotic offences that may fall within the loopholes of current legal systems. The chapter proposes a pragmatic approach, in order to define whether new robotic offences have to be established in addition to that which several scholars argue in the fields of international humanitarian law and current laws of war.