The regulation of militarised artificial intelligence

Abstract

According to the International Court of Justice (ICJ), the legal principles of the law of armed conflict (LOAC) apply ‘to all forms of warfare and to all kinds of weapons, those of the past, those of the present and those of the future.’ 4 One of the most difficult and complex discussions on the application of international law to novel technologies is that of autonomous weapons systems (AWS). For years, this discussion has been unfolding on the pages of academic journals and in the halls of Palais des Nations in Geneva, where the Group of Governmental Experts on Lethal Autonomous Weapons Systems (GGE on LAWS) regularly convenes. 5 This Group has not only greatly contributed to the sophistication of the debate on these weapons, but it has also facilitated a dynamic exchange of state positions on the scope of international legal rules and their application to emergent military capabilities with varying degrees of autonomy. The Group has adopted, by consensus, 11 Guiding Principles, the first of which affirms that ‘[i]nternational humanitarian law continues to apply fully to all weapons systems, including the potential development and use of lethal AWS.’ 6 While there is agreement that international humanitarian law 7 applies to all weapons systems, the difficult question that remains is how exactly it applies to those systems. This question of ‘how’ is related to the scope of the rules of international humanitarian law, including the precise contours of the duties imposed on parties to armed conflict. The aim of this chapter is to provide an overview of state positions on the regulation of autonomous military capabilities relying on AI and to outline three obligations of particular importance in ensuring the safe deployment of weapons with autonomy in decision-making: the obligation to conduct legal reviews of new weapons, the obligation to take precautions in attack, and the obligation to take precautions against the effects of attacks.