Artificial intelligence, autonomous weapon systems, and command responsibility under the law of armed conflicts

Recent developments in artificial intelligence technology have revealed the possibility of using artificial intelligence in weapon systems, and this has brought some concerns. The basis of this concern lies in the fact that target selection and attack decision, which is claimed to be in the monopoly of humans, will be transferred to machines with the emergence of autonomous weapon systems (AWS), and the ongoing problems of explainability and unpredictability of learning artificial intelligence systems. Although there is no AWS in use yet, artificial intelligence is used in many areas in the armed forces, but the problem areas prevent further use. AWS is a weapon system, not a combatant, in the context of the law of armed conflict (LOAC). Therefore, in the absence of specific regulations for the AWS, the weapons law and the general principles of LOAC should be applied. An analysis within the framework of the principles of humanity, military necessity, discrimination, and proportionality, which are the core principles of the LOAC, shows the possibility of legitimate use of AWS. In addition, as with other weapons, the responsibility of the operators and decision makers will arise from the violations that will occur as a result of the use of AWS. Therefore, there will be no responsibility gap. It is also possible to apply command responsibility besides individual responsibility. Command responsibility arose in the proceedings after World War the Second, later, it was included in Protocol No. 1 to the Geneva Conventions, and the concept became clearer with the proceedings of the international criminal courts. Today, it is accepted that command responsibility is a customary law rule. In order for a commander to be held responsible for the actions of his subordinates, there should be a command and control relationship and effective control, the commander should have knowledge of the crimes committed/about to be committed by his subordinates, and he should not have taken the measures within his power and ability to prevent, suppress and punish them. When these conditions are met, command responsibility can also be applied to war crimes resulting from the use of AWS.