The Morality of Autonomous Weapons: An Answer to the Threshold Objection

Autonomous weapon systems (AWS) can select and engage targets without human intervention. Proponents cite reductions in psychological, moral, and lethal harm as reasons in their favor. However, the threshold objection (TTO) predicts that, by making war less risky, AWS would lower the threshold for war, increasing the frequency of unjust warfare. TTO is noteworthy for two reasons. First, it turns the purported moral benefits of AWS into an argument against their use. Second, if sound, it suggests that AWS constitute an important exception to the weapons irrelevance principle, according to which a military’s choice of weaponry is irrelevant to jus ad bellum norms. In this article, I offer two responses to TTO. My first response suggests that the empirical conjecture underlying the objection – that AWS will increase unjust warfare – is unlikely to prove true and, even if it does, would equally threaten other military technologies. My second response demonstrates that even if AWS were to lower the threshold for war, this alone would not justify halting development or endorsing an international ban. The moral evaluation of AWS must weigh potential increases in unjust aggression against an expanded capacity for just interventions and a reduced level of civilian and combatant harm.