The proposal by Allen Buchanan and Robert Keohane to set up a Drone Accountability Regime (DAR) to help ensure better compliance with the existing law for lethal use of remotely piloted air systems (RPAS) is both welcome and timely. However, it is worth pointing out that using a drone as a component of a military operation does not automatically make that military action a “targeted killing.” Indeed, it is probable that much of the public concern and condemnation of drones is actually an objection to this type of attack, seen by many people as comparable to acts of assassination on foreign soil.Footnote ¹ Targeted killing is itself not a drone-specific concern, as such strikes can be carried out by missile, helicopter gunship, sniper, or a special forces team inserted into foreign territory. However, the perceived lower political cost of employing drones means that Buchanan and Keohane are particularly concerned about (1) their use in violation of sovereignty principles; (2) that they will be overused, creating an incentive to employ force as a first response instead of as a last resort; and (3) that they will be used to target civilian leaders and thus confuse, ignore, or undermine the distinction principle. The DAR is a response to these fears.

It is definitely worth taking a moment to reflect on the other ways that RPAS might be employed, and for which the DAR would presumably not be pertinent or required. Those whose lives have been saved by the accurate and timely delivery of close air support in a combat situation would certainly challenge the idea that drones are inherently “bad” and in need of a specific controlling regime. In recent conflicts drones have been deployed in mixed grouping alongside other platforms and assets to complement them rather than replace them.Footnote ² For the person who is taking incoming fire and is calling in air support to come to his rescue, he may or may not know if it is going to be delivered by a fast jet or an RPAS loitering overhead. Indeed, as long as the support arrives when required and is on target, those on the ground calling it in are very unlikely to be concerned where it has come from. In such circumstances, the idea that this specific military platform should require a distinct regime to govern its use would probably be greeted with some bemusement.

While it is of course possible for any weapon system to be employed without due care and attention, decisions to deploy munitions in the Afghan theater were governed by the same process regardless of which platform they were launched from. Unless acting in self-defense (in which case no further authority is required), any use of lethal force had to be authorized by an appropriately empowered engagement authority. In practice, this is normally the commander on the ground, who, having gone through and satisfied both the target eligibility procedures and collateral damage assessments, passes instructions through a qualified Forward Air Controller (FAC) to whatever platform is required to deliver the munition, be that an artillery battery or a drone.Footnote ³ However, while the FAC channels the request, “everyone involved in prosecuting a target, from planning through to pulling the trigger, has the obligation to comply with the Law of Armed Conflict and to reject any order that is deemed unlawful.”Footnote ⁴

As far as accountability to the law is concerned, one could argue that drone operators are among the most heavily monitored individuals connected to the battle space. They are covered by the same rules as everyone else, but like the pilot of a fast jet, “every movement of the joystick, every frame of camera footage and therefore every decision—or, indeed, hesitation or omission—made by the operator is recorded and can be pored over at great length following any incident.”Footnote ⁵ Unlike some other types of situations, there really is nowhere to hide from a bad decision in this particular environment.

It would also be churlish not to acknowledge the enormous advances in discrimination and proportionality that such new assets make possible. Indeed, Buchanan and Keohane do recognize some of these advantages, acknowledging that they can be more discriminate than other forms of violence. The physical separation of the operator from combat can offer advantages in terms of accuracy, proportionality, and discrimination. The fact that the operator is not directly at risk when on a mission provides them with some very real advantages over a person who is physically in harm's way. They have the ability to “step back and have a bit more of a Hamlet moment, as it were . . . you can hopefully double-check what you're doing is correct.”Footnote ⁶ This detachment provides time and space for decision-making that is, quite literally, a world away from the experiences of soldiers on the ground or even of pilots in hostile airspace. Given all their advantages and the fact that they put fewer people in harm's way on both sides of the equation, there may even be a moral imperative to employ more drones rather than fewer if they can do the same job as a manned platform.Footnote ⁷

The authors are correct when suggesting that the lack of friendly “body-bags” associated with using drones means that the military option may become more attractive when previously it might have been ruled out. This is a valid concern, and relates to the potential lowering of the threshold to the use of force. However, if there is a lower political cost in using such weapons, in some situations this might actually be a good thing. Imagine if we chose to intervene this time and prevented another Rwanda from happening. However, even given all these caveats—and potential benefits—large-scale humanitarian intervention is clearly not the type of military activity that Buchanan and Keohane have in mind. Nor are they concerned with the many other areas of contemporary drone use, from search and rescue missions to the provision of real-time intelligence of a situation. They are focused on targeted killings. And while their proposal is called a Drone Accountability Regime, it might be worth qualifying the title to make it clear the particular area of drone use that is being regulated.

Targeted killings arouse very strong emotions, and the sense of outrage is only increased when they are carried out in someone else's sovereign territory. That drones are relatively easy to deploy in this type of role is precisely where the DAR comes in. Such concerns are not new. Cicero records how the Roman Senate refused to countenance the assassination or “treacherous murder” of even a powerful enemy—one who was waging an unprovoked war.Footnote ⁸ Historically, there was a certain unease associated with the use of “underhand” tactics. After all, one picks out a victim rather than a genuine adversary in either an ambush or an assassination. As I have observed elsewhere, killing under such circumstances appears “unsporting or even cowardly.”Footnote ⁹ Nothing has really changed in attitudes today. For example, U.S. Attorney General Eric Holder was very clear when he argued that the current drone strikes against targets in Yemen and Pakistan were categorically not assassinations, insisting: “They are not, and the use of that loaded term is misplaced.”Footnote ¹⁰

Targeted killings are therefore rationalized and justified using very specific language.Footnote ¹¹ For example, rather than being a politically-inspired “hit,” or a revenge or punishment attack, a targeted killing—authorized only at the highest level and therefore removed from more routine situations—refers to “the intentional killing of a specific civilian or unlawful combatant who cannot reasonably be apprehended, [and] who is taking a direct part in hostilities, [with] the targeting done at the direction of the state, in the context of an international or non-international armed conflict.”Footnote ¹² Even though Buchanan and Keohane are specifically concerned with attacks upon civilians rather than military personnel, and the subsequent erosion of the core principle of distinction, such a definition can still be commensurate with the core idea found in all codifications of basic human rights: that “while there is an absolute right not to be arbitrarily deprived of life, the use of lethal force against an individual can still be justified if it is absolutely necessary for the defense of another person from an act of unlawful violence by that individual.”Footnote ¹³ If it were possible to arrest the individual or apprehend him in another way, clearly the attack could not qualify as “absolutely necessary”; but a terrorist leader who has directed attacks in the past and is planning to do so again has made himself liable to harm even if he has never picked up a gun himself or worn a military uniform. He may also be well protected and constantly moving from safe house to safe house in contested or hostile territory, making him impossible to apprehend in a normal way. In such a situation, if a state can act with a drone strike or some other similar method to prevent ongoing incidents of harm from such an individual, if the evidence is extremely clear, and if authorization is granted by someone suitably high up in, or even above, the normal chain of command, that would be a case of targeted killing carried out in self-defense—not an assassination.

But despite all of the high-profile denials and specific definitions combined with high-level assurances that the rules of drone use are scrupulously adhered to, there is still a problem. There is very rarely a public explanation for an attack, and without such transparency there is no way for anyone (outside the chain of command) to know the difference between an extrajudicial execution and a legitimate act of national self-defense. If it looks like an assassination, how is anyone to tell the difference? This leads us directly to the next problem: What happens when drones are used by those who do not follow such exacting procedures? After all, at some point they might be used by those who we would not consider “nice people.” When such actions are carried out in someone else's territory, this issue becomes even starker. Without any public explanation of why strikes are being carried out, what is to stop attacks on apparent noncombatants from happening more frequently, riding roughshod over the principle of sovereignty at the same time?

U.S. Attorney General Holder explained in 2012 that an attack across an international border would only take place “with the consent of the nation involved, or after a determination that the nation is unable or unwilling to deal effectively with a threat to the United States.”Footnote ¹⁴ But, again, this seems to ask as many questions as it answers. What does “unable or unwilling” actually mean in practical terms? What type of activity counts as a “significant threat”? If Russia were seeking to protect itself from an alleged dangerous terrorist but refused to share its evidence on security grounds, thus making the United States unwilling to act on Russia's request for support, would that make a Russian precision strike in Chicago acceptable?Footnote ¹⁵ Presumably not, but without more clarity it is difficult to see why the situation would be treated so differently.

As shown above, when used in a military theater alongside other military assets, drones or RPAS are treated in the same way as any other weapon system, governed by the same rules and, if anything, are actually subject to a higher level of oversight and accountability within the military organization deploying them. However, when drones are employed to carry out a targeted killing, particularly extraterritorially, the context does appear to be different, and this is where there is a real and desperate need for more transparency if the rules are not to be unintentionally eroded.

While enthusiastic about the Buchanan and Keohane proposal, I understand that the political challenges to implementation are likely to be profound. The chances appear slim that the United States, United Kingdom, or many nonstate actors, for that matter, would voluntarily subject themselves to additional extraterritorial accountability and acknowledge that an ombudsperson has the right to hold them to a standard and “judge whether they have fulfilled their responsibilities in light of these standards, and to impose sanctions if they determine that these responsibilities have not been met.”Footnote ¹⁶ Realists would question why anyone would accept the neutering of such an enormous asymmetric advantage over nearly every potential opponent by setting up a regime that limits the state's ability to act with freedom. However, as the authors rightly note, any current asymmetric advantage is inevitably going to be only a fleeting one as others continue to develop drones of their own. To give just one example, the Iranians unveiled their “Karrar” bomber drone back in August 2010, with President Ahmadinejad saying that the new plane would serve as a “messenger of death” (adding reassuringly that its key message was one of friendship).Footnote ¹⁷ Therefore, moving on this issue now with an eye to the near future would seem to be in the best long-term interests of those who currently have this capability. They will not merely be binding their own actions but will be establishing a norm that others may feel they are also obliged to follow, thus protecting everyone.

Recognizing that things will not happen instantly, Buchanan and Keohane describe a “process through which ‘first-movers’—in this case states that are already using lethal drone technology—could begin a multi-staged process of agreement-making.” This seems prudent and sensible. Even without the formal regime in place, one of the first steps could be to encourage more transparency when drones are employed in targeted killings. In the absence of a credible explanation, without such transparency a targeted killing might as well be considered an assassination or just plain murder—even if the action has been carried out with the best of intentions and is based on impeccable reasoning related to self-defense and the preservation of life in the face of an imminent threat. “Trust me, I am a defense professional” is not enough in this situation. How are others to follow and adhere to a norm that is effectively kept secret?

Obviously, full public disclosure in many situations is unlikely to be realistic. There will inevitably be security concerns attached to the release of certain types of sensitive information. However, the targeting actor can still give a clear idea as to the nature of the threat that the targeted person represented, how imminent the threat was, and so on, as far as is possible, so that the legal rationale can then be understood. If a state is not prepared to make that information available, then, as I have previously argued, we should refrain from calling such an action a targeted killing and instead call it what it effectively becomes—an execution. As this becomes the norm, the idea of making the more detailed evidence available to a trusted ombudsperson for independent scrutiny and verification may naturally develop.

To step forward and accept additional oversight, scrutiny, and accountability while one is in the technological lead greatly increases one's moral capital and makes it harder for others not to follow. As Philip Alston's report to the UN Human Rights Council suggests, genuine transparency combined with genuine accountability may ultimately be the best safeguard.Footnote ¹⁸ Buchanan and Keohane's proposal is a very good move in this direction.