With the use of artificial intelligence, so-called autonomous weapons systems could one day operate freely, choosing and eliminating targets without human involvement. With the seemingly inevitable proliferation of these systems, some will eventually violate international humanitarian law. Consequently, the establishment of legal responsibility for potential war crimes will be seriously impeded by the absence of a human in the command chain. Therefore, the individual responsibility regime under IHL would in many instances fail to hold a military commander or manufacturer of an AWS responsible for its wrongful actions. However, since the regime of state responsibility under the IHL framework does not require a mental element, it is preferable to that of individual responsibility, in order to avoid accountability gaps for the actions of AWSs. A strict responsibility regime for AWSs should furthermore be established to assure that states are held responsible for any damage cause by their AWSs.