The Legitimacy and Accountability for the Deployment of Autonomous Weapon Systems under International Humanitarian Law

Recent advances in automated weapon systems and decreasing need to have human operators in the loop of decision-making processes, has given the floor to one of the most controversial matters in the history of warfare: the emergence of Autonomous Weapon Systems, known as AWS. AWS it has been the source for significant controversy and heated international debate. One of the many reasons that are certainly behind this highly intense debate is the lack of a common ground for discussion.1 NGO’s, scholars, States struggle to find a common definition of AWS, to agree in their nature and legitimacy under International Humanitarian Law; and, finally, to contemplate the possibility of a ‘responsibility gap’ caused with the deployment of AWS. The focus of this research is to discuss the legitimacy and accountability for the deployment of AWS under International Humanitarian Law. The first endeavour was to prove that, first of all, it is necessary to organise the speech developed around autonomous systems in order to establish an AWS legal definition and to conclude that they should not be seen as ‘new agents’, but as more sophisticated ‘weapon systems’, looking at the structure of the embedded machine learning algorithms. The second endeavour was to prove that there are no reasons to proclaim a ‘responsibility gap’. AWS will require not only a higher level of responsibility from commanders, designers, programmers and technicians but also the distinction between different causes of violations of IHL caused by AWS (malfunctions, accidents and errors). The situation of a ‘accidents’ will require that International Criminal Law accepts the category of dolus eventualis, and in the situation of ‘errors’ States shall accept their full responsibility but in no manner is it possible to argue that IHL does not provide the tools to deal with the advent of warfare autonomous systems.