Moralities of Drone Violence

Front matter

Christian Enemark

For Bree and Henry

MORALITTIES OF DRONE VIOLENCE

Contents

Acknowledgements

List of Abbreviations

1 Introduction: Armed Drones and Drone Violence

Chapter 1 INTRODUCTION: ARMED DRONES AND DRONE VIOLENCE

Qasem Soleimani died in Iraq on 3 January 2020 when the Iranian general’s vehicle was struck by two missiles on a quiet road near Baghdad International Airport. With authorisation from the then-president of the United States, Donald Trump, the missiles had been launched by the remote operators of a MQ-9 Reaper drone aircraft flying overhead. Nine other people were killed in the attack: four Iranian officers and five members of the Iraqi Popular Mobilisation Forces.' US intelligence officials had earlier managed to obtain secret details of Soleimani’s visit to Iraq,” enabling the timely positioning of the Reaper. Then, for the first time, an armed drone was used by one state against a high-level official of another state on the territory of a third state. An international killing had occurred without any need for the killers to leave the United States, and without any immediate risk of their being discovered, captured and questioned by foreign authorities. Consequently, it has remained unclear what those US-based drone operators understood about what they were doing and what kind of violence they were perpetrating on this occasion. Was this drone strike against a foreign general from an unfriendly country an act of war, or was it an instance of ‘use of force below the threshold of war’?® Was it intended, in the manner of an intervention by armed police officers, to prevent an imminent and violent threat to public safety inside Iraq? Was the strike rather a means of punishing someone who was guilty of committing crimes against Americans? Or was it none, or somehow all, of those things?

The rationale offered by the US government at the time was a mixed one.’ An initial Defense Department statement described its drone-based killing as a ‘defensive action to protect U.S. personnel abroad’ whom Soleimani was ‘actively developing plans to attack’.’ Such language suggests that the strike was driven by military concerns and had a warlike quality. Yet the statement also claimed that Soleimani was ‘responsible for the deaths of hundreds of American and coalition service members and the wounding of thousands more’,® suggesting that the strike might also have been an act of punishment for past wrongdoing. Later, President Trump characterised the drone strike as being both preventive and punitive: a ‘bold and decisive action to save American lives and deliver American justice’.’ Regarding the need to act preventively, there was some official disagreement about whether Soleimani had been about to order an attack. US Defense Secretary Mark Esper acknowledged that he had not seen specific evidence that American embassies were in jeopardy immediately prior to the drone strike.* But Secretary of State Mike Pompeo insisted that Soleimani really had posed an ‘imminent’ threat, such that the president’s ‘decision to remove Qasem Soleimani from the battlefield saved American lives’.’? Even so, the punitive rationale for the drone strike remained a consistent feature in the US government’s explanations. Attorney General William Barr opined that ‘this concept of imminence is something of a red herring’ anyway, and Trump tweeted likewise that ‘it doesn’t really matter because of [Soleimani’s] horrible past!’.!°

Some US allies publicly supported the killing of Soleimani as an exercise in ‘self-defence’, although Iraq’s prime minister condemned it as an ‘assassination’.'” Then, in August 2020, this drone strike became the centrepiece of a report by Agnés Callamard, the United Nations (UN) Special Rapporteur on extrajudicial, summary or arbitrary executions. Callamard found it difficult to frame the strike as an action in war such as would attract the application of international humanitarian law (IHL). At the relevant time, neither the United States nor Iran appeared to consider that a condition of war existed between them,” and Callamard questioned whether ‘IHL standards’ were the best ‘fit’ for the purpose of legally assessing the known circumstances surrounding Soleimani’s death outside of an active conflict zone.'* The better view, she argued, was that international human rights law (IHRL) remained applicable to this instance of non-war killing." In which case, the US government was responsible for arbitrarily violating Soleimani’s right to life when it sought to obtain ‘justice’ in an improper way.“ Critically, though, Callamard also acknowledged that this was an example of how ‘[s]ome drone strikes’ raise ‘genuine uncertainty’ about the basis for their justification."

Moral uncertainty surrounding the use of armed drones more generally has been a persistent problem during the last two decades. This is partly attributable to the technological novelty of uninhabited aircraft equipped with video cameras and weapons, but ethical controversies mainly arise from the way in which some drone-using states have chosen to exploit and apply this emergent technology. In the associated debate among ethicists, lawyers and policy practitioners, the moral rectitude of particular drone strikes and of drone strikes in general has often been fiercely contested. However, these debaters have sometimes talked past each other, ethically speaking, because they have brought different conceptual understandings to bear upon the violence that is under consideration. In response, Moralities of Drone Violence aims to provide greater clarity, for debate purposes, by exploring and ordering a variety of ways in which the violent use of an armed drone can be judged as just or unjust. Accordingly, this introductory chapter outlines the book’s organisation of moral ideas around a series of concepts of ‘drone violence’: warfare, violent law enforcement, tele-intimate violence, and violence devolved from humans to artificial intelligence (AI) technologies. At the outset, though, it is important to establish what ‘armed drones’ are, how they are used, and who uses them.

ARMED DRONES: USES AND USERS

To begin, it is worth emphasising that the term ‘drone’ is employed here only for the sake of linguistic simplicity and because it is in common usage. The term is sometimes criticised (mainly by individuals with a military background) for being imprecise or carrying ‘negative connotations’. However, in a discussion of ethics that proceeds without prejudice, it is not clear that any analytical benefit would accrue from routinely replacing ‘drone’ with a more cumbersome term (such as ‘uninhabited aerial vehicle’ or ‘remotely piloted aircraft’). Arguably, then, it is better to employ simpler language and to assume that the ethical significance of drones lies less in what they are called and more in how they are used.

With the advent and advancement of drone technology, some tasks that could previously be performed only by the on-board pilots of aircraft can now be performed more easily by the remote controller of a drone. Moreover, the availability of drones has made possible the performance of new kinds of tasks. This is because, when an aircraft operates without a human on board, there is more scope for that aircraft to vary in size, and it can fly higher, faster and for longer periods than the human body can normally tolerate. Throughout the world, these technical advantages are mostly applied for benign purposes, and most drones are not armed. For example, remotely controlled and camera-equipped aircraft (sometimes with additional load-bearing capacity) have been useful in the provision of emergency healthcare,” the delivery of medical supplies,” the monitoring of endangered wildlife,” and the rescuing of pets.” Nevertheless, ‘drones’ in general can still seem fearsome and untrustworthy, even to people in need of drone-based assistance, because of the notorious association of some drones with violence.” This is mainly attributable to sustained media coverage of long-range strikes conducted by the US government since late 2001 (when the Afghanistan War began), although public attention has more recently turned also towards other uses of armed drones by numerous other governments.

For present purposes, an ‘armed drone’ is defined narrowly: it is a reusable aircraft that does not carry a human operator, carries at least one weapon, and incorporates on-board sensor and communication technologies for directing that weapon. This book is therefore not directly concerned with non-aerial (land-based or maritime) vehicles, unarmed drones that provide targeting assistance for other weapon platforms,” or with non-recoverable aerial devices (including missiles and so-called ‘kamikaze drones’) that crash into a target.” Each of these technologies is worthy of attention from an ethical perspective,’ but they are excluded here in the cause of pursuing a focused and coherent inquiry into the moralities of drone violence.

Uses for armed drones

An armed drone collects information from the environment where it is flying and delivers a weapon towards a target (on the ground or in the air) within that environment. For the most part, based on existing technology, such violence manifests in ground attack mode (involving air-to-surface weapons). When operating in this mode, a drone might be useful for protecting or supporting friendly or allied ground troops in armed conflicts. This could involve, for example, coordinating missile strikes with troop attacks and providing air cover to facilitate tactical troop withdrawals.” Alternatively, a drone that launches air-to-surface weapons might be used for killing a particular individual located within or outside a conflict zone, or perhaps for the domestic policing purpose of neutralising a criminally violent threat to public safety.

Two types of drone aircraft, controlled over a satellite radio communication link, have been used most extensively by the US government for striking targets on the ground in distant foreign territories: the MQ-1 Predator and the newer MQ-9 Reaper. The Predator is a 27-foot-long aircraft that can operate on 24-hour missions, up to 770 miles away from its base, at an altitude of up to 25,000 feet. The US Air Force (USAF) started using it for surveillance and reconnaissance in 1996, and in 2002 this type of drone was armed with two AGM-114 Hellfire missiles. These could be guided towards ground-based targets by the Predator’s on-board ‘Multi-Spectral Targeting System’ which integrates an infrared sensor, high-definition video cameras, and, for targeting purposes, a laser designator and laser illuminator.” The USAF eventually retired the Predator and transitioned to using mainly the MQ-9 Reaper,” which was designed from the start to carry weapons. The Reaper carries a similar set of sensor and munition-guidance technologies as the Predator, but it can fly farther and higher than its predecessor and is large enough (at 36 feet long) to be more heavily armed (with up to eight Hellfire missiles or with a variety of larger bombs).*

The salient feature of these large and long-range drones (as the US government has used them) is that they enable the killing, in near-real time, of another person made visible at a precise location in another part of the world. Indeed, in using the satellite-linked Reaper to project power abroad, the US has arguably been able to transcend limits of time and space to an unprecedented degree. However, smaller drones can be useful too, when operating in ground attack mode at short range, even though their payload capacity and flying time are much less than a Reaper’s.*’ In general, a small drone is operated by a single person within ‘line of sight’ range, which is the distance (actually extending beyond visual range) a control signal can travel between the airborne aircraft and a ground controller’s radio.** Examples of small-sized armed drones include the Tikad (equipped with a machine gun),” the Cerberus GL (for carrying a single bomb, a 12-gauge shotgun or a launcher for three grenades),*® and the Smash Dragon (which can fire assault or sniper rifles as it hovers over a target).” Or, for the dispersal of ‘less-lethal’ munitions such as pepper spray balls (perhaps as a law enforcement tactic), the Skunk Riot Control Copter is available.**

The ability to use weapons in air-to-surface mode is largely dependent on dominating the airspace. To this end, some drone manufacturers and governments are interested in armed drones that can use air-to-air weapons and prevail in aerial combat with other aircraft. The propellerpowered Reaper (made by General Atomics) used by the US and UK governments has been useful for conducting strikes against technologically weak adversaries on the ground in places where airspace is largely uncontested. However, this drone is not as fast and agile as modern fighter jets (like the F-22 and F-35),* so its utility is limited against the sophisticated airpower assets of more powerful adversaries like Russia and China.” Drones designed to operate only in ground attack mode could be modified, for greater ‘survivability’, to carry air-to-air missiles (like the short-range AIM-9 Sidewinder), but this might not be enough.*! In which case, according to some drone-using states, there is a mounting need for drones that are designed afresh to be swifter, stealthier and more heavily armed than their predecessors. Prototypes of drones that are thus more suitable for aerial combat include General Atomics’ jet-powered MQ-20 Avenger,

Northrop Grumman’s X-47B, BAE’s Taranis, and Dassault Aviation’s nEUROn.”

In the design and experimental development of such drones, there is a capacity for some of their system functions (such as navigation and mid-air refuelling) to be performed ‘autonomously’ (by an AI technology rather than a human).* Such a capacity is especially advantageous for a drone flying at extreme distances from its ground control station. This is because it can otherwise take several seconds for data and control signals to transmit via satellite between a drone aircraft and its remote human controller. This ‘latency’ has required the US government, for example, to reduce the risk of crashes by arranging for take-offs and landings to be handled by operators deployed close to the drone’s runway on a base abroad.” In general, reducing human involvement and reliance upon long-distance communication would accelerate function-performance processes, making some elements of a drone’s operation more efficient. However, it remains to be seen whether or how future drone-users will be able and willing also to incorporate AI into the process of controlling a drone’s weapons.

Users of armed drones

In the short history of armed drones, they have predominately been used by states. Violent non-state actors have occasionally equipped drones with air-to-surface weapons. In 2016, for example, members of the Islamic State group modified small, short-range, commercially available drones to drop grenades on enemy personnel, equipment and buildings in Syria and Iraq.* Some other incidents, though, have been misreported as ‘drone’ attacks by non-state actors when, in fact, the attacks involved a remote-controlled and explosive-laden aircraft being crashed into a target.*° State actors, by contrast, have tended to be much more serious and successful as users of armed drones, because they are far more likely to possess the technology, resources and infrastructure necessary to sustain this method of violence. Granted, most states are incapable of sustaining drone use on the same scale as the United States has done.“ Even so, the very idea of engaging with enemies on a remote-control basis remains powerfully attractive to some political leaders, and this adds impetus to the international proliferation of armed drones. The desire for these weapon systems might in some cases be attributable to a perceived need to project power in a new way. Or, as some authors have argued, acquiring armed drones might also be a response to domestic political pressure and a ‘post-heroic’ aversion to casualties among national personnel.** In any event, around the world, drone technology has continued to attract government interest and investment, either as a means of supporting personnel deployed to dangerous places or as an alternative to such deployment.

In retrospect, it is now possible to identify a ‘first drone age’, beginning in late 2001, during which armed drones were used only by Israel, the United Kingdom and the United States. For the Israeli government, armed drones (built in Israel) have mainly been used for conducting short-range airstrikes in nearby territories.*” The UK government, using US-made drones, has conducted strikes inside Afghanistan, Iraq and Syria.’ And the US government, with its global network of military bases and intelligence resources, has been able to deploy its drones in those places too, as well as in Libya, Pakistan, Somalia and Yemen.”' For about fourteen years, this tiny group of drone-using states retained exclusive access to armed drones. Then, in 2015, there began what has since been described as a ‘second drone age’, characterised by the emergence of new drone manufacturers and the widespread proliferation of weaponised drone technology.”

In September of that year, the Pakistani army announced that it had used one of its Burraq aircraft to conduct its first drone strike, killing three ‘high-profile terrorists’ inside North Waziristan (a region in north-western Pakistan).°? Then, in early 2016, a Nigerian military crew operating a Chinese-made Cai Hong drone conducted a strike, against members of the insurgent Boko Haram group, in a remote forest in north-eastern Nigeria.” Soon afterwards, Iran began using its armed Shahed 129 drones to strike targets in Syria.” And, in the same year, the Turkish government began using a new type of armed drone (designed and built in Turkey), the TB2 Bayraktar.° Turkish drones have since been used extensively within the country’s borders against the separatist Kurdistan Workers’ Party, and also against Islamic State and Kurdish groups in northern Iraq and Syria.” Moreover, Turkey (like China) has become a major exporter of armed drones for wartime use by foreign governments.°®

In 2017 a Syrian government drone fired at US-led coalition forces operating inside Syria,” and two years later both sides in Libya’s ongoing civil war began deploying foreign allies’ drones against each other. While the UN-recognised Government of National Accord (GNA) in Tripoli used Turkish-supplied Bayraktars, the rival Libyan National Army was supported by the United Arab Emirates’ use of Chinese-made Wing Loong drones against GNA targets.® France carried out its first-ever drone strike in late 2019, using an armed Reaper in support of French commandos under attack by ‘a group of terrorists on motorbikes’ in central Mali.“ In 2020, during fighting over the Nagorno-Karabakh territory (claimed by both Armenia and Azerbaijan), the Azeri military reportedly used imported Bayraktars to conduct missile strikes.* In 2021, Ethiopia’s national government purchased armed drones from Turkey and Iran for use against the rebellious Tigray People’s Liberation Front. And, in 2022, the Ukrainian government’s defence against invading Russian forces included the use of small drones carrying small bombs as well as, again, missile-armed Turkish Bayraktars.™

At the time of writing, according to a database maintained by the UK-based organisation Drone Wars, fourteen states have conducted drone strikes, five states are involved in exporting armed drones, and fourteen states have imported them. In addition, fifteen other states are ‘near’ to operating armed drones.® For example, Italy has received approval from the US government to arm its two Reapers with Hellfire missiles and laser-guided bombs.” India has ordered ten armed Heron TP drones from an Israeli manufacturer.” And Germany’s parliamentary defence committee has reportedly approved the purchase of 140 Herons as part of a boost to the country’s military capabilities following Russia’s invasion of Ukraine. Thus, it is clear that drone proliferation is real and ongoing, and that many lives will be at stake if this leads to a worldwide increase in the violent use of drones. However, the future extent and effect of this violence are arguably not determined by technological advancements and political factors alone. Rather, that future might yet be shaped also by concerns about justice and by associated commitments to exercise principled restraint. Accordingly, the remainder of this chapter turns to the making of moral judgements about ‘drone violence’, which is open to be conceptualised in various ways.

DRONE VIOLENCE: CONCEPTS AND MORALITIES

The academic literature addressing the use of armed drones has grown to incorporate a wide variety of disciplinary perspectives, but it also exhibits a common tendency among authors to refer to such activity as ‘drone warfare’. Much of the time, ‘warfare’ is indeed what is going on when a drone is used. In some circumstances, though, it is impossible, implausible or insufficiently helpful to conceptualise drone use in that way. For example, one drone-user might rather engage in what is essentially a law enforcement action. Another instance of violent drone use might be difficult to recognise as being either warfare or law enforcement,” or it might appear somehow to bear the characteristics of both.” In addition, the concept of ‘drone warfare’ arguably does not fully capture the essence of an individual’s hands-on experience of using an armed drone, and nor does this concept account well for the future possibility of such use being controlled to some extent by AI technologies. For these reasons, the task of conceptualising the use of armed drones needs to begin, instead, with a more general term: ‘drone violence’. This then makes available a range of approaches to thinking ethically about violent drone use (extending beyond the Just War morality traditionally applicable to war violence), which in turn affords greater scope to discern and address its potential to generate injustice.

With these objectives in mind, it is firstly worth noting that the word ‘violence’ has the analytical advantage of not being a euphemism for something else. Also, it is important to emphasise that that word is employed here in a morally neutral fashion (without carrying a negative connotation). That is, Moralities of Drone Violence necessarily rejects a ‘legitimist’ definition of violence that ‘incorporates a reference to an illegal or illegitimate use of force’. No assumption is made that violence per se is, by default, ‘something we should disapprove of’,” so the act of referring to ‘violence’ should not be understood here as one of condemnation. Rather, in this book, ‘drone violence’ serves as a basic term upon which a number of more specific concepts (to be introduced later in this chapter) can be overlaid. Thus, the discussion of multiple moralities can proceed based on drone violence being conceptualised as ‘warfare’, ‘violent law enforcement’, ‘tele-intimate violence’ or ‘devolved (to AI) violence’.

Conceptualising violence, then judging it

For present purposes, conceptualisation is understood as something that logically precedes ethical assessment. That is, prior to considering whether and why a given instance of violence is just or unjust, it is necessary to establish what kind of violence this is. It is sometimes the case in practice, though, that value judgements are smuggled into the conceptualisation process, causing ‘the conceptual and the ethical’ to be ‘muddled together’. And, typically, such muddling originates in a person’s preformed desire ultimately to render a justification or a condemnation. The moral judgement (for or against) that they would prefer to reach about an instance of violence is what then drives and prejudices a choice about framing it conceptually. However, this judgemental ‘conceptualisation’ is clearly not intended to reveal the pre-ethical nature of whatever violence is under consideration. Instead, it is a technique for enabling that violence to escape, or to be captured by, certain moral expectations (and any associated rules and governance measures).

On the one hand, for example, to label a killing an ‘assassination’ is not to conceptualise violence but to condemn it. This is because, around the world, there is general agreement that assassination is (and should be) illegal; there cannot be ‘good assassinations’. On the other hand, referring to violence as ‘war’ can be part of an attempt to afford it some (potential) legitimacy, for there can be ‘just wars’. As Ian Clark has observed, the war concept thus often ‘serves as a residual . . . normative source in its own right’.”* And, according to Emily Kalah Gade, ‘war’ can be a political term used by state actors attempting to ‘legitimize and formalize the violence in which they engage’.” Conversely, the condemnation of violence sometimes consists in describing it as ‘not war’, such as when circumstances of extreme asymmetry are perceived as morally wrong. This is the moral judgement that appears to drive Jean Baudrillard’s claim that the 1991 Gulf War, involving overwhelming US airpower and few American casualties, ‘did not take place’. And, similarly, the huge imbalance in casualties among Israelis and Palestinians caused during Israel’s Operation Cast Lead in Gaza (launched in December 2008) is what prompted Avi Shlaim to describe this violence as ‘not a war... but a one-sided massacre’.” In both cases, an alternative approach would have been to conceptualise the relevant violence (as ‘war’, for example) and then to judge it (as unjust war).

This approach, arguably, is also the better one when it comes to ethical assessment, because it is less vulnerable to prejudice and more conducive to meaningful debate. As different ways of thinking about violence ‘profoundly shape the possibilities and limits of the ethical debate that can take place’, settling on a concept first is important if debaters are to avoid talking past each other. If it is accepted that war is going on, a debate in the language of military ethics can proceed. But if the violence under consideration is essentially a practice of law enforcement, the ethical debate about this is necessarily restricted to the moral vocabulary of criminal justice (qua policing or punishment). A prior debate can still be had about whether a condition of war or law enforcement exists in the first place, but this is a conceptual debate not an ethical one. So, for example, the answer to Michael Gross’s question of whether ‘targeted killings’ are ‘[permissible] acts of self-defence or [impermissible] extrajudicial execution’ depends upon ‘which of the two paradigms that justify lethal force — war or law enforcement — we choose to analyze targeted killing’.*’ The same approach, applied to the use of armed drones, means that expectations about what conduct is appropriate ‘are dependent upon the particular concept of the use of force in which [drones] are initially located’.

The conceptual location of state violence within either of two paradigms — war or law enforcement — is critically significant,*’ because the morality that attends the latter (a peacetime activity) is traditionally much less tolerant of killing than is war’s morality. It is generally understood that many of the violent actions which warriors may commit in war may not be committed by other state agents (police and executioners) when violently enforcing the criminal law. The difference, then, between abiding by the more permissive ‘ethos of warfare’ or the more restrictive ‘ethos of policing’ can be measured in terms of who may be killed and in what circumstances.** And, as between these different pathways of departure from every human’s presumptive right to life, the humanitarian outcomes (measured in injuries and deaths) are likely also to differ accordingly. In matters of law enforcement, the ethical expectation is that state violence is directed towards a person based only on wrongful behaviour (what they are doing or have done). By contrast, wartime targeting is permitted also according to a person’s generally threatening status (as a combatant or direct participant in hostilities). Consequently, there tends to be far more moral scope for killing enemies in war than there is for killing criminals in peacetime.*°

This traditional notion of distinct moralities of violence is rejected by some moral philosophers who insist, instead, that there is only ‘one morality’.*” According to Jeff McMahan, for example, the justifications for killing people in war are no different from those that apply in other contexts. So, rather than focusing on the supposed moral difference between war and peace as overarching conditions, his ‘individualist’ approach to the assessment of violence emphasises the moral inequality among participants in a violent encounter. Ethical assessment then comes down to distinguishing between, on the one hand, ‘those who are in the right’ who are ‘permitted to use force’ and, on the other hand, ‘those who are in the wrong’ who have no such permission.** According to this reasoning, a robber is clearly ‘in the wrong’ when they mug a pedestrian on the street. And, whereas the robber is not morally permitted to use violence for this unjust purpose, the pedestrian does have permission (as a matter of self-defence) to violently resist the violence being wrongly visited upon them. Moreover, a passing police officer who intervenes to protect the attacked pedestrian is morally entitled to act violently too (as a matter of other-defence).

In seeking to apply this one (individualist) morality to all violence, however, there is a practical difficulty: it is often much harder to establish who is ‘in the right’ in the politically messier reality of violent encounters that are called ‘war’. Notoriously, each side in a war tends to proclaim that they alone have justice on their side, so comparisons to righteous police and culpable robbers must struggle to achieve relevance. Also, as Michael Walzer has argued, war as ‘a coercively collectivizing enterprise’ makes it impossible to pay attention to ‘each person’s moral standing’

when inquiring into the overall justification for this kind of violence.’ Thus, it is generally considered more useful to follow the orthodox ‘collectivist’? approach: to regard war and peace as morally distinct paradigms, and to hold war-wagers and law-enforcers to differing standards of violent conduct.

Four moralities of drone violence

In this book, the idea of multiple moralities is critical because it enables a widening of the scope for normative inquiry into the use of armed drones. By conceptualising drone violence in various ways, different and novel approaches to moral judgement are made available, and this has the potential overall to reveal more about what it means for this violence to be just or unjust. To conceptualise drone violence more narrowly, though, is to run a greater risk of failing to capture the full range of potential injustices. On the matter of justification, there is no shortage of scholarship that examines the use of armed drones from a legal perspective.” And it is indeed possible to criticise or defend such use purely by reference to the international law on resorting to force, to IHL, and/or to IHRL.”! By contrast, in Moralities of Drone Violence, the meaning of ‘morality’ is understood to extend beyond points of legality.

Specific laws are occasionally mentioned herein as reflections of underlying moral principles, but sometimes the discussion of an ethical concept or a type of injustice needs to proceed in the absence of any corresponding legal rule. Also, for present purposes, any invoking of legality is not intended to truncate what it means to behave morally or to be affected by others’ immoral behaviour.” As Elke Schwarz has suggested, it would be short-sighted to conflate ‘thinking ethically’ about drones with merely ‘adhering to existing laws’.”? And it is worth recalling that law tends to work ‘externally’ in the sense that people obey it mainly because they believe ‘the law reflects some collectively held judgment of the right.” Yet, as Valerie Morkevičius has observed, morality has a broader and deeper meaning and influence: it occasions individuals also to be ‘inward-looking’, thus shaping behaviour based on ‘how we see ourselves’ and ‘even if no one is looking’.”°