AGI crimes? The role of criminal law in mitigating existential risks posed by artificial general intelligence

Abstract

The recent developments in applications of artificial intelligence bring back discussion about risks posed by AI. Among immediate risks that need to be tackled here and now, there is also a possible problem of existential threats related to Artificial General Intelligence (AGI). There is a discussion on how to mitigate those risks by appropriate regulations. It seems that one commonly accepted assumption is that the problem is global, and thus, it needs to be tackled first of all on an international level. In this paper, I argue that national criminal laws should also be considered one of the possible regulatory tools for mitigating threats posed by AGI. I propose to enact AGI crimes that complement the varieties of legal responses to existential risks that might motivate and speed up further regulatory changes.