Preventative Governance of AI in the Military Domain

Abstract This chapter argues that the established framework of international law requires states to engage in preventative governance of military AI systems across their lifecycle, and especially in the pre-deployment stages (ie before operational use). Drawing from international humanitarian law, international human rights law, arms control law, and the law of state responsibility, it interprets existing obligations to ensure respect for, and to prevent violations of, international law as supporting concrete standards of anticipatory regulation, design constraints, internal oversight, and external restraint. The chapter thereby outlines a forward-facing governance approach which is both legally required and practically necessary, in light of the increasing autonomy, scale, and speed of military AI systems.