Who is to blame for Autonomous Weapons Systems’ misdoings?

This Chapter analyses who (or what) should be held responsible for behaviours by autonomous weapons systems (AWS) that, were they enacted by a human agent, would qualify as internationally wrongful acts. After illustrating the structural problems which make ascription of responsibility for AWS’ activities particularly difficult, when not impossible, the alternative routes proposed to solve the ensuing responsibility gap will be assessed. The analysis will focus, in the first place, on the international criminal responsibility of the individuals who, in one way or another, are involved in the process of production, deployment and activation of the AWS. The possibility to hold the deploying State accountable for AWS’ wrongdoings will then be gauged. Subsequently, attention will be paid to the responsibility of the corporations manufacturing and/or programming the AWS. It will be observed that these options may solve some responsibility problems more effectively than critics of AWS are ready to admit. At the same time, it will be shown that, unless a no-fault liability regime is adopted, autonomy in weapons systems is bound to magnify the risk that no one may be held to answer for acts which are objectively in contrast with international legal prescriptions. Also, it will be argued that, given the complementary relationship among the various forms of responsibility under international law, proposals aimed at focusing solely on one of these at the expense of others are incapable of leading to satisfying results.