Toward a Drone Accountability Regime

Abstract

From a moral standpoint, lethal drones are intrinsically no worse as a means of warfare than bombing or sending commandos to kill enemies. From the perspective of their users, they have six major advantages over more conventional weapons: they are often cheaper; their use can be more readily concealed; they allow for more precise targeting, with the potential for less “collateral damage”; their use can involve less serious infringements of sovereignty than invasion by troops; and they may be less likely to provoke widespread hostile reactions by the population of the country in which they are used than military operations involving troops on the ground. But these advantages generate three major risks: of violating sovereignty, of over-using the military option, and of making it more difficult to identify violations of constraints against targeting noncombatants. To deal with these risks, a Drone Accountability Regime is needed that imposes obligations on states, which in turn would be required to impose them on their agents. Since it would be infeasible to negotiate a treaty-based legal regime at present and for the foreseeable future, the Drone Accountability Regime should be informal and should involve transnational actors as well as states. Its key principle should be transparency, helping enable civil society to hold states accountable, and its central agent would be an Ombudsperson with broad authority to investigate situations and publicize her findings. No institution can ensure that states, or operators, are held fully accountable to appropriate standards of conduct, but such a regime could increase the degree of accountability for the use of lethal drones.

States are using lethal drones unilaterally, without transparency or accountability, and will do so more in the future—unless there is an international regime for regulating them. We argue that such a Drone Accountability Regime is essential to achieve accountability, and we identify the main features of an appropriate regulatory regime that has the flexibility to evolve to meet new challenges and opportunities. The first part of this article explains special risks of lethal drone use that call for new regulation, and the second part argues for an international regime incorporating transnational actors as well as states. The regulatory regime we propose does not carve out an exception to the laws of war; rather, if implemented, it would help ensure better compliance with the laws of war in the case of lethal drone use.

We begin by discussing how international law applies to lethal drone use. Here we adopt what has been called the “war paradigm” for dealing with issues of terrorism, rather than the “policing paradigm.” Terrorist attacks are more like acts of war than like ordinary crime. They have broad political purposes and are typically targeted at groups. Large-scale attacks have been carried out by well-organized groups with modern weapons and often advanced military training, rather than by individuals seeking personal advantage. Moreover, it is often impossible to capture terrorist suspects for trial, especially when operating in failed states or lawless regions. The policing paradigm, by contrast, is designed for generally well-ordered societies that suffer from occasional and small-scale violence engaged in by individuals or small groups (typically in the hopes of personal advantage). Even advocates of the policing paradigm admit, as did Kenneth Roth in a well-known 2004 article, that it only applies where a law enforcement system exists. ¹

No actual conflict perfectly fits either of these ideal types, and policing activity plays some role in countering terrorism. Nonetheless, the large scale of major terrorist attacks means that the war paradigm is a better fit than the policing paradigm for the sorts of conflicts that make a regulatory regime for lethal drone use valuable. The war paradigm has another distinct advantage: There is a well-developed body of law governing it, and there is an impressive consensus that this body of law is authoritative. The same cannot be said for the policing paradigm. The regulation of policing activities varies considerably in different states, and there is no body of international law for policing that is comparable to the laws of war. Consequently, we will address the issue of regulation on the assumption that the uses to be regulated fall under the war paradigm and the legal and moral norms that this paradigm entails. ²

We specify three major risks of lethal drone use, which we frame in terms of widely accepted principles of just war theory that are already encompassed by the humanitarian law of war or by UN Charter–based international security law. These risks all derive from two of the very advantages that make drones so appealing to policymakers: lower costs and difficulty of detection. ³ We therefore rest our normative argument on the least controversial elements of international law.

The second part of this article explains the necessity of a global regulatory regime for lethal drones. It then proposes an institutional design for such an international regime, which has two levels of accountability. At the global level, individual states will be held accountable to interstate and transnational institutions in the Drone Accountability Regime. At the national level, states will themselves hold their own lethal drone operators accountable.

Since we view a treaty-based international regime as infeasible under present circumstances, we instead sketch an informal global regime that could accomplish some of the goals of such a formal regime and make acceptance of a more formal regime more likely. Our informal regime includes states, nonstate actors, and an Ombudsperson with broad authority to investigate and publicize potential abuses. Such a regime is designed to enhance accountability through transparency and publicity and, in countries where this is feasible, to mobilize domestic constituencies to support compliance with the regime. It is supposed to work both ex ante, by requiring that states specify appropriate procedures in their decision-making processes for all drone strikes; and ex post, by requiring public justifications for specific strikes after they occur. We emphasize the need for dynamic accountability: that is, the capacity for ongoing reassessment of the regulatory environment and modifications to it.

This Drone Accountability Regime would focus principally on states: that is, territorial entities that are recognized as such by other states, typically through acceptance of their membership in universal international organizations such as the United Nations. Of course drones could be deployed by nonstate actors as well, or improperly used by their state-employed operators, but only states have the authority to implement effective regulations. States would therefore be responsible for holding their own operators accountable through internal procedures, and for controlling nonstate actors operating from their territories.

We argue that it would be infeasible at the present time to prescribe detailed rules concerning specific strikes, because states would be unlikely to countenance such a restriction on their sovereignty and because different procedures could be appropriate within different political systems. We do not foresee circumstances under which nonstate actors that use violence—and that might at some point obtain lethal drone technology—could be included in an international regulatory scheme. Instead, the goal is to create norms of use by states that meet minimal rule-of-law standards. ⁴

We realize that it will be very difficult to reach agreement on a Drone Accountability Regime, even in the informal, non-legalized form that we propose; and we expect that others will improve on our proposal, which we believe is the first in the literature. We also hope that policymakers will take the issue seriously, so that when the time is ripe, action can be taken. Even though it will be difficult to negotiate such a regime, we think that in principle it should be feasible, given that the major drone-deploying states have incentives, as in arms control regimes involving nuclear weapons or missile technology, to limit both the proliferation of these weapons and their unregulated use by others. We therefore outline 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. We then explain the incentives that both these first-movers and later participants would have for participating in the process.

Drone Use and International Law: Attractions and Risks

Three main bodies of international law apply to the use of lethal drones. UN Charter–based international security law chiefly addresses lawful recourse to war. The humanitarian law of war addresses permissible weaponry, legitimate targets, and the treatment of prisoners, of wounded combatants, and of civilians and property in areas controlled by armed forces during war. International human rights law governs the use of force, especially by states, outside the context of war. When taken together, these three bodies of law supply impressive doctrinal resources for regulating lethal drone use. In this section, we rely on these bodies of international law, but organize our discussion around the attractions, and especially the risks, of lethal drone use.

For present purposes, we define lethal drones as unmanned aerial systems: robotic aircraft capable of inflicting death on human beings. ⁵ We will suspend judgment about the permissibility of the use of fully autonomous lethal drones because we do not know enough about their capabilities. ⁶ Instead, we will focus on lethal drone use that involves continuous human control.

Attractions of Lethal Drone Use

Lethal drones are employed instead of other, less invasive coercive measures, such as the use of aircraft, cruise missiles, or relatively small teams of special operations forces. Compared to these alternatives, four features of drone technology encourage its use. First, they are often cheaper to use. Second, drones may make it possible to kill targets that might otherwise be immune from attack, especially since they can “loiter” over a target for hours and can respond immediately to commands. Third, from a purely technological standpoint (setting aside for now problems of user error and abuse) lethal drones are capable of greater precision in targeting, with less collateral damage to persons or property than conventional weaponry, especially since they can be diverted by the systems operator at the last moment. ⁷ Fourth, they function without exposing human weapon operators to the risk of grievous injury, death, or capture.

Three Risks

The chief special risks of lethal drone use by states are (1) violations of sovereignty by unauthorized incursions into another state's territory; (2) overuse of the military option, encouraged by lower costs, especially in terms of reduced risk to the user's military personnel and the avoidance of negative reactions to ground troops; ⁸ and (3) less easily detected violations of the discrimination principle. These three risks (discussed in greater detail below) can all be framed in terms of relatively uncontroversial moral and legal principles. Nonconsensual intrusion into a state's territory is a violation of a fundamental principle of sovereignty that is common to international ethics and international law. Unnecessary recourse to military force is a violation of UN Charter–based international security law, the humanitarian laws of war, and of mainstream just war theory; and unnecessarily killing (rather than capturing) is a violation of the humanitarian law of war and of the moral prohibition on excessive force, as well as a violation of the human right against arbitrary deprivation of life. Targeting noncombatants is a violation of the discrimination principle, which specifies that only those individuals engaged in violence or who actively support violence may be legitimately attacked, and that efforts must be made to prevent lethal harm to bystanders insofar as possible.

Recent evidence suggests that drone killings have violated the discrimination principle much less than other forms of violence directed against terrorism, but the number of civilian deaths from drones is not negligible. ⁹ As we will argue below, breaches of the discrimination principle through drone use are often more difficult to identify than are breaches resulting from bombing campaigns or the use of ground troops.

Drone use also raises questions about the distinction between legitimate and illegitimate human targets. In traditional just war theory and in current humanitarian law of war, deliberate killing of civilian leaders is prohibited. Insofar as civilian leaders may have made the decisions for violence in the first place, this prohibition is not grounded on a distinction between the innocent and the guilty, but is, on the contrary, instrumental in character. First, targeting civilian leaders may diminish the prospects for a negotiated settlement, since it is these leaders who are typically authorized to make peace. Second, distinguishing between civilian and military leaders helps to maintain the supremacy of civilian leadership. If civilian leaders are targeted, military leaders may thereby attain unchecked authority.

In “irregular” warfare, however, formal roles are less well defined and the distinction between military and civilian leaders is often arbitrary. When it is impossible to draw a non-arbitrary distinction between civilian and military officials, all leaders who are directly involved in the planning or execution of military operations are legitimate targets in principle, although for pragmatic reasons a state might decide to maintain the fiction of a distinction in order to maintain lines of communication with its adversaries. ¹⁰

In his speech at the National Defense University on May 23, 2013, President Barack Obama addressed many of these questions. ¹¹ He defended the necessity of drone use to combat terrorism in states, including in parts of Somalia and Yemen where state control is “tenuous”; he pointed out the danger of sending U.S. troops into areas controlled by al-Qaeda or other hostile forces; he asserted that such drone use is effective; and he claimed that it is less likely to engender backlash than sending in American special forces. “Our operation in Pakistan against Osama bin Laden cannot be the norm,” he declared. He also admitted the risks of civilian casualties in drone strikes and expressed deep regret for them, although he argued that these risks were less severe than if the United States instead used other means of military force. With respect to oversight, he declared that “not only did Congress authorize the use of force, it is briefed on every strike that America takes.” In many respects we agree with President Obama, particularly in seeing drone strikes as less bad with respect to likely casualties than the alternatives. But we conclude that, given the three risks noted above, even his thoughtful defense of U.S. drone policy is not sufficient.

1. First risk: Violating sovereignty. Under international law, sovereignty includes the right of a state to be free from unauthorized incursions by other states into its territory, whether these involve soldiers or unmanned military devices. Given the fact that states are the primary protectors of the well-being of individuals and groups and the most important institutions for establishing justice and the rule of law, there are powerful moral as well as legal reasons to observe this norm of territorial sovereignty. But if a drone can penetrate a state's borders and destroy a target without leaving a “signature” that links it to a particular state, it may become less costly to violate this norm.

If states A and B are formally at war with one another, the issue of consent to such territorial intrusion is irrelevant. If no such war exists, then a state planning to deploy drones within another state's territory is obligated to seek its consent; if it acts without consent, it violates the norm of territorial sovereignty. Sometimes, however, there may be no functioning government from which to obtain consent. This might be the case in the midst of a civil war in which no one faction persistently controls most of the territory. In such cases, and with respect to failed states—those that are in a condition of anarchy, without even a minimally functioning government—no agent may be qualified to give consent. In such situations there can be no obligation on the part of a state seeking to deploy drones there to try to obtain prior consent.

Admittedly, what constitutes a failed state will be contested. Many states lack effective control over their entire territories and populations; indeed, states that are well functioning are unlikely sites for drone use by outsiders. Pakistan, for instance, is clearly not a failed state although it is a troubled one: its government operates in much of the country, and there is a political process at work to determine who has official authority. Yemen is somewhere near the margins: there is a government whose writ runs, more or less, in part of the country, but large areas are effectively beyond its control, and its rule is periodically threatened by insurgent groups. ¹²

It could be argued that even where the state has failed there is a kind of residual sovereignty residing in the people of the state. This idea best explains the intuition that it is impermissible for other states simply to annex the territory of failed states or impose their own favored form of government at will, without regard to the preferences of the population. ¹³ On this view, which is fundamental to liberal political theory, the rights of sovereignty do not attach, properly speaking, to the government or the state, but to the people, and can therefore survive the breakdown of political order. But the difficulty of making this claim operational is that in a failed state there is no political process through which the preferences of the people could be ascertained.

We conclude that nonconsensual strikes are permissible within the territory of failed states when transnational violence is being organized within their borders and valid state consent is impossible to attain, unless these strikes are (1) used as part of an unjust military operation or are (2) part of a larger policy project that would violate the residual sovereignty of the people.

There is another ambiguity regarding the consent issue. Some governments support the use of drones within their own territories, but are unwilling to admit this fact due to domestic opposition and their own weakness, with Pakistan and Yemen being prominent examples of this dilemma. In such situations, to require formal consent could thwart the desires both of the drone-using government and the state on whose territory the drones are used, and thereby fail to deter terrorism. But a system where the drone-wielding government can simply claim consent, without clear evidence that it was authoritative, is obviously subject to abuse. The Drone Accountability Regime that we propose below seeks to address this critical issue of consent. In particular, we argue that even with respect to failed states and other states that cannot feasibly give their explicit consent in advance to drone strikes, rigorous procedures to ensure ex post accountability are required.

2. Second risk: Overuse of the military option. Since lethal drones lower the costs of military interventions, states could be inclined to employ them when other methods, from coercive diplomacy to economic sanctions, might achieve the same goals. Although military force is not always a morally inferior option—sanctions that grievously harm innocent people may be worse—there are sound reasons in favor of a strong presumption against its use. When the costs of using military force are lowered, one must expect more of it.

Lethal drones also create incentives to kill rather than capture. For example, attempting to capture well-protected insurgent or terrorist leaders located in remote regions through the use of special forces is a high-risk endeavor. For one thing, there is a significant probability that the mission may fail. Moreover, there may be other high costs, such as troop casualties or negative responses to the presence of foreign soldiers in the area. The use of lethal drones can eliminate both of these risks. How much risk an agent ought to bear in order to capture rather than kill a target is disputed, but the lower costs of lethal drone use may tempt an agent to avoid any risk whatsoever.

3. Third risk: Less easily detected violations of the discrimination principle. Since terrorist and other revolutionary organizations are necessarily clandestine, it is not always clear who the members or affiliates of these groups are or how to distinguish them from civilians. Responding to such ambiguity, the United States has resorted to “signature strikes,” in which non-identified individuals are targeted if they fit a broad category, such as being members of al-Qaeda or the Taliban, or of associated or similar groups. ¹⁴ In a public defense of the Obama administration's drone policy, Harold H. Koh defined al-Qaeda–associated forces as: “(1) an organized, armed group that (2) has actually entered the fight alongside al Qaeda against the United States, thereby becoming (3) a co-belligerent with al Qaeda in its hostilities against America.” ¹⁵ But this definition is quite vague in the absence of information about how it has been defined in actual operations. For example, it is not clear whether “entering the fight” is interpreted broadly in the U.S. drone program to encompass individuals who are not actual combatants or are not directly contributing directly to violent activities. The key problem is that the processes in which these criteria are applied are hidden from view. In other words, there is lack of transparency.

Further, the risk of detection of violations by drones may be much lower than that of violations by more conventional bombing or attacks by soldiers. In conventional warfare, violations of the discrimination principle are often detected because of reports by “whistle-blowers” among the military force of the perpetrators, survivors among the enemy, humanitarian workers in the conventional battle zone, or aerial (including satellite) photography capable of identifying mass graves. Drones make all of these less likely. Reduced risk of detection therefore lowers the costs of use, and may thus cause drones to be deployed against people who are not legitimate targets. Just as the lower costs of drone strikes may encourage overuse of military options in general, the reduced risk of detection may encourage violations of the discrimination principle.

The Moral Acceptability of Drones and International Law

None of the preceding areas of moral and legal controversy is peculiar to the use of drones. Disagreements about the norms governing intervention in failed states or what actually constitutes consent arise with regard to all forms of forcible intervention, and especially with targeted killings. How to distinguish legitimate targets in cases of “irregular” warfare has been hotly debated at least since the beginning of the “war on terrorism” and complicates the normative picture for the use of force generally, not just for the use of drones. Similarly, the question of distinguishing between civilian and military leaders in irregular conflict has long vexed just war theory and would continue to do so even if drones had not been invented.

Drones have advantages for legitimate states in their struggles against terrorism, and we see little reason to believe that lethal drones pose such serious risk, compared to other lethal activities, that they should be banned. But the risks of drone misuse are real. In simplest terms: Current international regulation is inadequate, because actors who control lethal drone use are not held accountable by any existing international body for acting in conformity with relatively uncontroversial moral and legal norms that clearly apply to their behavior. We therefore propose below institutional measures to reduce the risks that we have identified. ¹⁶

Establishing a Drone Accountability Regime

Adequate regulation at the national level cannot be counted on without the stimulus of international regulation. At present, virtually all lethal drone use is conducted by the United States. And, as noted above, American drone use lacks institutionalized transparency. This lack of transparency and the possible abuses that accompany it have persisted even though the United States is a liberal democracy with a comparatively robust civil society. It follows, therefore, that there is even less reason to believe that national regulation can be put in place to provide adequate accountability in the case of less democratic states that are likely to acquire the capability to use lethal drones. ¹⁷

The inadequacy of a purely national drone-use policy is predictable since even ethical government leaders will often have strong incentives to misuse drones in ways that generate the risks discussed above. One incentive is that it can be politically costly even for an elected chief executive to stand up to the national security bureaucracy, which combines expertise, an institutional commitment to security without a comparable commitment to international law or due process, and high status in society and with democratic legislatures. Another incentive is that political leaders are strongly inclined to minimize harm to their own citizens, and hence to give insufficient weight in their calculations to the risk that their actions pose to foreigners—an incentive that may be particularly strong in liberal democratic states. One valuable function of the international humanitarian law of war, and perhaps even more emphatically of international human rights law, is to help correct for the understandable but nonetheless potentially damaging deference that democratic leaders are likely to show to their security bureaucracies, and to the bias that stems from purely democratic accountability. ¹⁸

Accountability

Accountability to a broader than national constituency is central to our proposal for a Drone Accountability Regime. The term implies that “some actors have the right to hold other actors to a set of standards, to 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.” ¹⁹ Accountability is necessarily a power term: Those who can hold policymakers accountable exercise power over them. That is, they have the capacity to impose costs on the policymakers and thereby to increase the likelihood that the latter will respond to their demands. Since the efficacy of accountability depends on context, accountability needs to be dynamic. The relevant institutions must have the capacity to reevaluate over time the adequacy of the goals and the aptness of current standards for achieving them, and to revise both as needed.

In addition, the goals of a system of accountability should be explicitly distinguished from the standards established to meet those goals, for two reasons. First, the goals of an accountability system may be equally well served by different standards. Second, it is essential that the procedures to follow be clear to operators, so that they do not have to make judgments about whether certain actions will attain the specified goals, since these judgments are likely to be highly subjective and biased. To achieve the goals of accountability, therefore, it may be desirable simply to insist that operators follow specified standards of conduct: to use certain heuristics or proxies for norms of accountability. Furthermore, different standards may be suitable for different states. Achievement of the goals of accountability is therefore compatible with a plurality of standards of accountability.

Organizational Forms and Principles for a Drone Accountability Regime

We are now in a position to describe the organizational form that our accountability regime would take. In an ideally lawful world an international drone accountability regime would be treaty-based, with enforcement provisions and a permanent secretariat. However, it is difficult to envisage such a regime being negotiated by the major powers and put into effect at any time in the foreseeable future. For instance, the landmines regime—the result of intense advocacy effort—still has not been agreed to by China, Cuba, India, Russia, North Korea, South Korea, or the United States—in other words, the major actual or potential users of landmines. ²⁰ At the other end of the institutional spectrum would be a purely transnational organization, created by networks of civil society organizations seeking to promote voluntary norms of responsible drone use and transparency. However, such a transnational body would surely encounter serious operational difficulties. In particular, it is hard to imagine that it would have much impact on the policies of authoritarian states in which civil society actors are unable to operate freely. In the absence of an international drone regime, furthermore, it would be difficult for transnational networks to organize and focus their activities. ²¹ Without genuine state leadership, a regime to regulate lethal drone use would be ineffective.

Since a strong, formally constituted interstate drone regime seems infeasible and merely transnational arrangements insufficient, we suggest the formation of a Drone Accountability Regime (DAR) as an informal interstate arrangement. The clearest analog to the DAR is the Missile Technology Control Regime (MTCR), which was established in 1987 and now has thirty-four members, although China, India, and Pakistan have all declined to join. In 2002 all MCTR members except Brazil developed the Hague Code of Conduct Against Ballistic Missile Proliferation, seeking to establish norms of constraints on missile proliferation. Admittedly, the MTCR is focused only on proliferation, not on deployment and use, so is in some ways less sensitive politically than our proposal. Furthermore, the MTCR has no enforcement provisions of its own, and indeed its enforcement largely relies on the United States to sanction entities that violate MTCR standards. ²²

Our proposal for a Drone Accountability Regime is more ambitious than the Missile Technology Control Regime, and we recognize that for the sake of feasibility it might need to be scaled back, yet we believe that even a limited Drone Accountability Regime could make a real difference. The DAR would have an Assembly of States, represented by Permanent Representatives as in the United Nations, which would meet on a regular basis to consider issues brought before it; a Transnational Council of nonstate organizations, giving a voice to transnational organizations; and an Ombudsperson to investigate complaints and serve as an agent of accountability. Its rules would govern the use of drones, not just technology exports. It would not have legally binding enforcement powers, but its findings could impose reputational costs on states that flagrantly flout its provisions.

As the Missile Technology Control Regime experience suggests, authoritarian countries will be unlikely to join such a regime, at least initially. For such states to accept the Drone Accountability Regime, efforts by more democratic states to provide incentives to join would be essential. But even without initial participation by authoritarian states, a Drone Accountability Regime could reduce instances of abuse while establishing norms that might be generalizable more widely in the future.

Among states that enable civil society action, active involvement by civil society and transnational groups will be crucial for keeping the issue in the public view, rather than it simply becoming the province of the national security bureaucracy. That is, a Drone Accountability Regime will only have a significant impact if it also becomes the focal point for mobilization by transnational networks with roots in domestic politics. As the experience of human rights shows, such transnational networks can have an impact in some (but not all) types of countries, if they help to mobilize domestic constituencies. ²³ Civil society pressure on democratic governments will be especially important since leadership by democracies is key to the establishment and functioning of a DAR.

In addition to regulating state behavior, the Drone Accountability Regime will have to grapple with the use of drones by nonstate entities. Insofar as these actors are under state control, the doctrine of state responsibility applies. But insofar as lethal drones are used by nonstate actors not under state control, the DAR would provide an important institution for coordinating forceful action by states against such actors, with authorization from its Assembly of States. If, for example, a nonstate actor such as the Islamic State of Iraq and Syria (ISIS) were to acquire lethal drones, the DAR could authorize action to eliminate such drones as well as sanctions against entities that provided ISIS with this technology. Anticipating such situations, states might find the DAR potentially attractive as a focal point for coordination.

An Assembly of States is an essential element in a Drone Accountability Regime, since state authority would be necessary to make its provisions effective. We do not seek to specify voting arrangements in this initial proposal, since they are inherently dependent on political feasibility and since, due to the informality of the DAR, the Assembly would not make binding rules. Instead, it would operate as a forum for negotiation and bargaining, and a focal point for publicizing actual or possible violations of the agreed rules. The DAR would also have a Transnational Council of nonstate organizations. Such organizations as the International Committee of the Red Cross, which historically has played a pivotal role in developing the humanitarian law of war, would be invited to join the Transnational Council, help form its rules, and invite other members. The Transnational Council would be guaranteed wide access to information and would determine the rules for its own membership. The Council's function would be to maintain a high level of transparency and publicity, so that criticism of lapses from the provisions of the regime would be observed and publicized. Since the Drone Accountability Regime would be informal, however, none of its bodies would wield mandatory powers over states.

The norms that a Drone Accountability Regime would establish revolve around the concept of accountability, as we have described it above. States would be held accountable—to other states and to the Ombudsperson—for following regime norms, including the norm of only transferring lethal drone technology, under DAR rules, to other members of the regime. Drone operators would be held accountable not directly to the DAR (which would require an independent bureaucracy not likely to be acceptable to states), but to their own states, which in turn would be accountable to the members—state and nonstate—of the global institution. Such two-level accountability—international and internal—has two main attractions. It is more feasible than a rigidly specified set of rules that all states must follow in their internal arrangements, because it allows greater control for individual states, conferring on them a range of discretion as to which internal accountability measures they employ, so long as they meet the international institution's broader, largely procedural criteria for acceptable internal accountability. And it is more consistent with two realities: first, that at this early stage in the use of lethal drone technology no one is in a position to prescribe a uniquely optimal first-level accountability regime; and, second, that allowing some diversity in first-level accountability regimes, under appropriate conditions of transparency, can facilitate both state acceptance of a DAR and the identification of best practices. An international regime that allows constrained pluralism in internal accountability measures can create an environment for learning about how best to achieve and maintain accountability.

The key actor to ensure accountability would be an Ombudsperson, roughly modeled on the position created by the United Nations Security Council in 2009 to review claims regarding abuse of Security Council Resolution 1267. ²⁴ The Ombudsperson in the 1267 regime, Judge Kimberly Prost from Canada, has been quite effective in representing people who claim they have been wrongly listed as terrorists. She has successfully extracted information from governments, and as of 2013 had delisted twenty-six individuals, including Yassin Abdullah Kadi, who originally brought the crucial case before the European Court of Justice. ²⁵ The Ombudsperson would be selected by the Assembly of States in consultation with the Transnational Council, ensuring transparency and exerting pressure on participating states to make a publicly defensible choice. The Ombudsperson would be directly accessible by (1) individuals who claim that they or their associates have been wrongly targeted, (2) states with complaints against another drone-using state, (3) relevant NGOs, and (4) drone operators, or anyone else in the chain of command for drone operations, who has a complaint about a particular drone strike or the general procedures under which strikes are undertaken. Further, the Ombudsperson would have the right to publicize any complaints he or she receives. He or she would also have the right to initiate a process by which compensation could be awarded in cases where there is clear evidence that individuals had been wrongfully targeted. In short, the Ombudsperson would provide both a focal point for accountability and an individual (and staff) capable of interpreting often ambiguous actions and assigning responsibility for them.

The three major entities constituting a Drone Accountability Regime—the Assembly of States, the Transnational Council, and the Ombudsperson—would act through three key mechanisms designed to enhance effectiveness: (1) a plurality of initiatives, (2) multiple arrangements to ensure transparency, and (3) mobilization for dynamic accountability. All of these arrangements would rely, at the national level, on a template for regulatory supervision: that is, an explicit set of standards, linked to the goals of the regime. Each member state would adopt such a template and accept procedures, including review procedures, by which it would be held accountable for adherence to it. In particular, each state would establish a national supervisory body with explicit, public procedures for ensuring that the template for regulatory supervision accepted by the state is actually implemented.

If there were more agreement on standards, and less uncertainty about how the technology and tactics associated with drones would evolve, one might imagine that the Drone Accountability Regime would establish a single template, similar to various arms control regimes over the last half century. But in view of large differences in interests among states, and the pervasive lack of trust in the intentions of others, requiring a single template would most likely preclude creation of the regime. We therefore favor creating opportunities for a variety of initiatives, in recognition that a reasonable response to uncertainty is to establish an experimentalist regime. Among the more democratic members of a Drone Accountability Regime, experimentalist governance could provide for dynamic accountability through pluralistic initiatives, although “neither the success of deliberation nor the inclusion of all relevant stakeholders is assured.” ²⁶

With respect to these initiatives, any group of five states, or members of the Transnational Council from at least five different states, or the Ombudsperson could propose templates for the Drone Accountability Regime. The rationale for this proposal is to retain the principle of pluralism while forcing some coherence and interstate compromise, so that a limited number of distinctive practices results. Each such template would be registered with the Office of the Ombudsperson and provide a set of largely procedural internal accountability measures, designed to reduce the risks we have identified. As noted above, each of these measures would have to include the establishment of a national supervisory body, which would be accountable to the Assembly of States and whose actions would be monitored by the Transnational Council. Each participating state would have to adopt some such template and publicly announce its adherence to it, but the provisions of the various templates could vary considerably, meeting the principle of pluralism.

These measures will only be effective if civil society is mobilized to monitor state action and demand accountability. Drone use is a sensitive topic, and the national security structures of states tend to be closed and self-protective. For this type of global regime to work effectively, activists in civil society will have to mobilize and stay mobilized. ²⁷

As we have noted above, accountability implies transparency. One aspect of transparency is passive: everyone must be able to see which template each state has adopted. But the regime would also incorporate active transparency. On the model of the United Nations Human Rights Council, each state would have to defend its template, and its adherence to it, before other states in public. In the Drone Accountability Regime, the Transnational Council would be authorized to select up to ten nongovernmental actors to participate with the state members at these review sessions. These sessions should occur on a regular basis, at least every five years. The Ombudsperson would play an active role in publicizing actions and issues, and in maintaining an environment of transparency.

Transparency is an essential condition for the success of a Drone Accountability Regime, since accountability depends on it. In the end, however, no institutional provisions can guarantee the DAR's effectiveness, which will depend on the mobilization of transnational groups—in particular NGOs such as Human Rights Watch—and the maintenance of their ability to mobilize domestic publics.

Ex Ante and Ex Post Accountability Mechanisms

Each acceptable template for internal accountability would include two types of accountability provisions—ex ante and ex post—in order to structure the process by which lethal drone strikes are actually authorized and executed. If feasible, there should be ex ante authorization of drone strikes. Such authorization is infeasible in the case of failed states, and may be infeasible when the territorial state is too weak politically to provide such authorization; in such situations, ex post authorization must be relied upon. But when the territorial state explicitly supports such strikes and is willing to say so, explicit written ex ante authorization provides the strongest basis for its legitimacy.

In the absence of explicit ex ante authorization, the drone-using state must take a number of general public measures on its own to bolster transparency. If the drone strike is to take place within the territory of a failed state—that is, if there is no legitimate political authority to consent to the incursion into that territory, or if there is serious disagreement as to who the legitimate political authority is (as in cases of civil war)—the drone-using state must explain publicly ex ante why the military necessity of using drones in the failed state is sufficient to overcome a general presumption against the use of force without state consent.

Further, the drone-using state must publicize its targeting criteria, along with an explanation of how these comport with current humanitarian law of war criteria for distinguishing legitimate from illegitimate human targets. The publicized targeting criteria must also explicitly address the need to satisfy the other two major concerns of jus in bello theory and the laws of war: the proportionality requirement and the requirement of military necessity.

In addition, the drone-using state must issue a policy statement that allows drone operators to refuse to initiate or to countermand drone strikes that they believe may be in violation of any of these three basic principles. It must also provide evidence that this policy is made known to drone operators, and that drone operators who refuse to initiate strikes or who countermand them on these grounds will not be subject to inappropriate penalties and will have recourse to the Ombudsperson if they believe they have been inappropriately penalized.

Ex post provisions in the Drone Accountability Regime are as follows:

1. 1. In the absence of clear ex ante authorization, a lethal drone-using state must provide a plausible public rationale, within two weeks of the strike, explaining why it was unreasonable to fulfill the requirement of prior consent in that particular case. Furthermore, all members of the Drone Accountability Regime must commit publicly to participating in good faith in an inquiry, organized by the Ombudsperson, as to whether this excuse for not obtaining prior consent was valid.

2. 2. After each particular strike, the drone-using state must explain its targeting decisions. ²⁸ If challenged, it must provide credible assurances that the strike is not an element in a larger set of actions that would violate the right of the people of the failed state to preserve their territory intact or their right to determine their own form of government.

3. 3. The national supervisory body (which, as noted above, every participating lethal drone-using state must have) is required to conduct periodic reviews of all records generated through implementation of the above ex ante provisions and to forward summaries of these reviews to participants in the Drone Accountability Regime. As noted above, the DAR will conduct periodic examinations of each state's policies, in which process the Ombudsperson will play a prominent role.

4. 4. The national supervisory body of each lethal drone-using participating state is required to conduct periodic reviews of its ex ante accountability measures, and the process of review is to include independent members of appropriate epistemic communities and humanitarian and human rights NGOs as well as staff members from the Office of the Ombudsperson. The results of these reviews are to be publicized and made available to participants in the Drone Accountability Regime.

5. 5. Drone-using states are required to keep written records of each lethal drone strike, including estimates at the time of the strike of its effects (number of casualties, killed versus wounded, legitimate targets killed or wounded versus “collateral damage”), to seek confirmation of the initial reports over time by gaining additional information, and to note and attempt to explain any discrepancies between initial and later reports. Such reports are to be provided to the Office of the Ombudsperson, which will distribute them to other participants in the regime.

The Issue of Feasibility

We acknowledge that our proposal for a Drone Accountability Regime is ambitious. But such a nonlegalized accountability regime as the one we have outlined would be a great step forward. The informal regime that we suggest will not be easy to create and may or may not be feasible under current circumstances. It would have to overcome opposition from organizations and politicians in the United States who do not wish to sacrifice any advantage from the fact that the United States is, for the time being at least, by far the leader in drone technology and deployment. And our proposed procedures for transparency will be resisted, as they always are, by authoritarian states as well as by national security bureaucracies everywhere.

We emphasize that we are not proposing the construction of a legalized, binding regime. Such a regime would almost surely be politically infeasible. Very few global, treaty-based regimes have been constructed since the World Trade Organization went into force in 1995, and it is particularly difficult to create formal international regimes on issues involving security. The national security bureaucracies of all states that rely on the unilateral use of force, including the United States, China, Russia, and Israel, will surely be skeptical of the DAR proposal. So we are aware that our proposal probably pushes the limits of political feasibility.

Critiques to the effect that the DAR would forestall attempts to create a much stronger, legally enforceable regime are therefore, in our view, based on utopian premises. They overlook the basic fact that states that now use drones or expect to acquire this capacity will resist binding legal constraints on their actions just as they have for landmines, which is a much less valuable technology. Those who would oppose our informal Drone Accountability Regime because it does not represent their ideal, legalized solution would be sacrificing the good for the unattainable.

Fortunately, we can identify some incentives to form or join an informal regime such as the one we have proposed. Among states that are accustomed to using force unilaterally, the United States stands out as a democracy with a vigorous civil society and without a bitter territorial dispute, such as that between Israel and the Palestinians. Furthermore, the United States, as the first-mover, has incentives to try to shape the rules of the regime while it still has some leverage to do so, deriving from its superior technology and deployment capabilities. By taking a leadership role, the United States could also gain reputational benefits and encourage other states to join the regime before these states build up large drone-oriented military units of their own. ²⁹

As with the Missile Technology Control Regime, other states will have incentives to join in order to please the United States and in hopes of some particularistic benefits, as well as to further the overall objectives of the regime. More broadly, the citizens of a liberal democracy like the United States should support an international institutional approach to achieving accountability for lethal drone use for two reasons: to ensure that other states will be held accountable, and also to enlist international institutional resources to make sure that accountability provisions in the United States are working properly. If civil society groups were mobilized, some democratic countries would have domestic political incentives to promote the Drone Accountability Regime. For other, often less democratic states, more self-interested benefits will be crucial.

Some of these benefits will be specific to the drone regime. As noted above, one requirement of the DAR should be that members can only transfer drones or drone technology to states that are members of the regime, giving an incentive to states that wish to purchase drones to join the DAR. In addition, members would gain more detailed information about what other states are doing, derived from access to information within the context of the drone regime and greater access to nonlethal drone technology. Other benefits of membership would be reputational. Also possible—but more difficult to arrange since they depend on cross-issue linkages—would be such extraneous benefits as the possibilities of trade deals or aid from the United States or other leading powers.

Even on the most optimistic scenario, it is difficult to imagine states that rely on the unilateral use of military force (with the exception of the United States) to agree to a Drone Accountability Regime. China, Russia, and Israel would therefore be unlikely candidates for membership. But forming a DAR without these states would nevertheless generate some reputational benefits for states that did join, as well as establish a set of norms that even nonmembers might seek to emulate to avoid opprobrium. Yet the level of likely resistance means that it would be entirely unrealistic to expect such a Drone Accountability Regime to appear suddenly. ³⁰ States will need to move step-by-step, assuring themselves that other states are also implementing effective national-level regulation on drone use. As in so much of world politics, reciprocity—through a tit-for-tat strategy—is critical to cooperation. ³¹ The negative side of reciprocity is that if some states with major drone capability refuse to join the regime, those who committed to the regime early may reserve the right to reduce their commitments, to withdraw, or even to dismantle the regime. The fact that the regime we envisage would be informal would make such conditional steps more feasible.

Yet one can imagine a virtuous spiral. First-movers, such as the United States, would have to take credible steps to induce others to reciprocate; but one can imagine a pattern of positive reciprocation, leading to stronger templates being accepted by states. Best practices could emerge that come to be valued by many constituencies as normatively valuable, enhancing incentives even for self-interested states concerned mostly about their own reputations to embrace them. If adherence to drone-related norms occurs in a gradual and transparent fashion, states can be assured that they will not be disadvantaged by the general move toward more systematically controlled drone use. Over time even an informal regime could gain credibility and specificity, adding more specific agreements going forward, increasing positive incentives to conform to it, and also increasing the costs of noncompliance.

Initially, some states joining the regime may do so because it is informal and they regard it as likely to be ineffective—a form of “cheap talk.” Such pseudo-adherence is not all bad. Indeed, such a perception of soft inefficacy could make it more feasible to form such a regime, even as it puts barriers in the way of its immediate efficacy. Over time, however, this expectation of ineffectiveness could be falsified as a result of vigorous civil society mobilization and the consequential generation of reputational incentives to follow the standards established in serious templates. Nonetheless, it must be emphasized that such progress would not be likely to occur as a result of formal legal measures. On the contrary, civil society must be active. No set of formal institutional arrangements can guarantee success. All that these arrangements can do—since for political reasons they will not involve legal obligations—is provide focal points for activity by civil society groups, putting pressure both on recalcitrant states and on those states that could be required, for effectiveness, to underpin the regime with sanctions.

Our proposal could have value even if never implemented, since it would focus attention on the deficiencies of current lethal drone use—in particular, the lack of transparency and accountability. Such attention could stimulate others to produce a better proposal, or at least add to the pressure on the U.S. government to improve its current policy on drone use. In any event, it is important to have a proposed Drone Accountability Regime before us as a model of an institutional arrangement to which we should aspire, and against which we can evaluate actual arrangements.

Max Weber said that politics is “a strong and slow boring of hard boards.” Establishing a Drone Accountability Regime would not be easy, and making it effective would be even more difficult. But establishing a DAR is a desirable goal and therefore worthwhile as a subject for discussion. Such discussions, furthermore, could help catalyze the civil society movements that would be needed to create it. In this way, the process of deliberation on a Drone Accountability Regime would itself be a valuable, if small, step toward a better institutionalized world order.