Much of the debate over the moral permissibility of using autonomous weapons systems (AWS) focuses on issues related to their use during war (jus in bello), and whether those systems can uphold the principles of proportionality and distinction. This essay, however, argues that we ought to consider how a state's portended use of AWS in conflict would affect jus ad bellum principles, particularly the principle of proportionality. The essay argues that even the clearest case of a defensive war against an unjust aggressor would prohibit going to war if the war was waged with AWS. The use of AWS to fight an unjust aggressor would adversely affect the ability for peaceful settlement and negotiations, as well as have negative second-order effects on the international system and third party states. In particular, the use of AWS by one state would likely start and arms race and proliferate weapons throughout the system.
Contents I. Introduction II. Jus ad Bellum III. Proportionality and AWS "We have to conduct battles without any contact, so that our boys do not die, and for that it is necessary to use war robots."--Dimitry Rogozin, Deputy Prime Minister of Russia (2)
"The DoD of FY2011-2036 Unmanned Systems Integrated Roadmap observes that warfighters continue to value the inherent features of unmanned systems, especially their persistence, versatility, and reduced risk to human life."--United States Department of Defense (3)
The above sentiments about saving troops' lives underlie much of the debate and perceived advantages of lethal Autonomous Weapons Systems (AWS). AWS are armed weapons systems, capable of learning and adapting their "functioning in response to changing circumstances in the environment in which [they are] deployed," as well as capable of making firing decisions on their own. (4) While it is certainly rare to see a fully candid statement regarding the desirability of saving one's own troops, such sentiments are among the motivating factors for the development and deployment of AWS. (5) Most scholarly arguments over the legal and moral permissibility of lethal AWS, however, do not cite these purely self-interested reasons, but focus instead on whether AWS are capable of upholding jus in bello principles, particularly the principles of distinction and proportionality. Proponents often cite that machines will be better able to distinguish between combatants and noncombatants than human soldiers. Furthermore, since machines are not affected by emotions, they will refrain from engaging in retributive acts against civilian populations. (6) Detractors argue the opposite, citing the difficulty of telling the difference between combatants and noncombatants, particularly in counterinsurgency wars. Moreover, they argue that such AWS will be unable to fight proportionately because the judgment required for such calculations is beyond the programming and learning capacities of these systems. (7) They claim that in bello proportionality is context-dependent, and artificial intelligence now and for the foreseeable future is unable to make such difficult judgment calls. (8) There is some discussion of the importance of risk to human warfighters, but this mostly concerns the impact of AWS on aspects of military virtue. (9)
This debate is an important one and should not be dismissed lightly. However, I would like to ask another question about the moral permissibility of AWS, namely, whether AWS is compatible with jus ad bellum. In particular, is it permissible to wage war knowing that it will be fought with these systems? I argue that AWS pose a distinct challenge to jus ad bellum principles, particularly the principle of proportionality. I will attempt to show that even in the case of a defensive war, we cannot satisfy the ad bellum principle of proportionality if we knowingly plan to use lethal autonomous systems during hostilities because of the likely effects on war termination and the achievement of one's just causes.
My argument has two parts. First, I briefly outline ad bellum principles and offer the clearest case of a just defensive war. Second, I argue that ad bellum proportionality calculations require us to look beyond a one-round game with an enemy force to the overall consequences of a war. In particular, one must not merely look to the counterfactual of not fighting, but what the war as a whole will look like as well as to the probable consequences that will result. In this way, I follow Thomas Hurka's conception of proportionality, arguing that it is inherently tied to all other ad bellum principles. (10) Thus if one cannot satisfy proportionality, the other principles will fall, like a line of dominos, along with it.
JUS AD BELLUM
Jus ad bellum is traditionally comprised of six principles: just cause, right intention, proper authority, last resort, the probability of success, and proportionality. Just cause provides the relevant normative reason to wage war, such as self-defense or defense of others, while right intention prescribes the proper reasons for acting. For instance, I may be unjustly threatened with war, thereby providing a just cause, but engaging in "defensive" war with an eye toward colonizing or annexing the aggressor state's territory is clearly impermissible. One's intentions in fighting must match one's publicly announced or official reasons for fighting.
The principle of proper (or sometimes legitimate) authority states that only legitimately recognized authorities may declare war. In the middle ages, this was clearly the province of princes. Today, it is the province of sovereign states. (11) One prominent criticism of the principle of proper authority comes from the direction of non-state actors. International law, for example, does not accord any privileges to these actors, but does place obligations on them. (12) While it is certainly true that states can wage war against non-state actors, the just war tradition would not consider these non-state actors proper authorities. (13)
Last resort and the probability of success are rather straightforward. Last resort requires that states attempt all reasonable alternatives available, such as diplomacy, arbitration, or countermeasures, before resorting to hostilities. (14) Probability of success, likewise, requires states to assess whether an actor is able to achieve its just cause through fighting. If a state is unable to do so, it would be futile to bear the costs of war. (15)
Finally, we have the principle of proportionality. Ad bellum proportionality, as distinct from its in bello counterpart, requires that we look to the overall consequences of a proposed war. (16) We must engage in a forward-looking counterfactual analysis and compare "the war and its effects to what the world would have been like had the war not occurred." (17) This calculation is notoriously difficult. Indeed, scholars like Hurka argue that such calculations cannot be "made simply or mechanically," for not all harms are of equal weight to potential benefits, and not all benefits are permitted in one's analysis. (18) For instance, we cannot count the overall economic benefits of engaging in a war, for those goods are unconnected to the relevant goods of waging war. According to Hurka, the "relevant goods" are those that proceed directly from, or are contained within, the just causes. (19) I will expand my discussion about proportionality throughout the remainder of this essay, but for now, I will follow Hurka in assuming that we must weigh only the relevant goods of waging war against all of the foreseen evils of doing so.
Let us turn now to the classic case of the defensive use of force. Unfortunately, it is not within the scope of this essay to address third party defense, harming of bystanders, and lesser evil justifications. (20) For our purposes, I will examine only a case in which an unjust aggressor (State A) intentionally and without justification threatens the central rights of another state (State D), namely the rights of territorial integrity and/or state sovereignty. (21) Under a forfeiture theory of self-defense, State A loses its right not to be harmed by threatening an imminent violation State D's rights. State D may inflict harm on State A to thwart an attack against it and to potentially restore State D's rights. The harm inflicted on State A must be necessary and proportionate. As noted above, only those benefits related to the just cause of defense will count in an ad bellum proportionality calculation, but all foreseen harms are included.
In the classic defensive scenario, most international lawyers, politicians, and some just war theorists would claim that State D is permitted to use proportionate force to thwart this attack. (22), however, would like to press on this intuition. In particular, in the case of an unjust threat, is a state permitted to respond with AWS? To restrict the case, let us assume that State D would contemplate sending only AWS and not a combination of AWS and human troops. This assumption prohibits us from bootstrapping additional...