Ex Machina

Over the last two months this blog looked at some of the developments in the realm of drone warfare and in some cases the greater use of drone technology. In the following, the key debates from these developments will be highlighted. 

First off, as stated in this blog’s introduction, it is important to recognize how we got to this point. Drone technology has been a long way coming, but looking at its history, its arrival seemed far from inevitable for most of the 20^th century. Today’s remote warfare could only come about due to the emergence of digital technologies coinciding with a new geopolitical paradigm. Experimentation with UAVs in the pre-digital era was largely aimed at taking pilots out of conventional military flights in combat or reconnaissance scenarios, limited by high costs and low reliability. It was not until the War on Terror was declared, that armed drones for targeted strikes saw substantive use. Not only had digital technology made the promise of remote strikes via drone more feasible, the new era of asymmetrical warfare under US-dominated unipolarity also opened up legal and political spaces of opportunity. Armed drones are as much a result of CIA counter-insurgency strategies as they are weapons of war and this shapes their applications until today. As such, it is important to recognize that the topic of drones is never a purely technological one, but equally a legal, political and ethical one. 

This complexity leads to, among other things, military drones playing a split role in public consciousness. While, as the historic overview and the widespread use of drones show, they represent great achievements in the minds of military leaders and intelligence agencies, their use, particularly outside of active warzones, continues to be accompanied by denials and lack of information. Documenting the extent, shape and consequences of the US drone campaigns continues to be immensely difficult task for journalists and NGOs. In fact, the US government has only recently taken steps back on transparency in drone use outside of warzones. A complex web of journalists, official voices and on-the-ground witnesses is required to keep tabs on even basic facts, such as the number and identities of those killed by drone strikes. And it is just this type of information that could inform a debate about the viability, legality and ethics of military drones. 

These debates are arguably vitally important now, as drone technology continues to be in constant flux and its uses are likely to increase as more and more countries acquire drone capabilities. Recent developments, such as the announcement of the Agile Condor project, which aims to allow drones to act more autonomously in identifying targets, suggest that the ethical questions we are facing today could have serious implications in the foreseeable future. The use of unmanned aerial vehicles today already represents a significant step towards the automation of war, when compared to conventional military flight in the late 20^th century. The tendency toward greater automation and more autonomy for machines seems to be ongoing and while fully autonomous drones are far from a certainty in the foreseeable future, the possibility should likely inform current debates about policy, military protocols and legal regulation. 

These debates often seem hampered by only looking at combat drones in the form in which they exist right now. Drone warfare to this day is largely marked by being very unilateral. This has changed slightly with terrorist and insurgent groups starting to use commercially available drones, but the possibility of a conflict with drones on both sides seems barely present in the public debates in drone-operating countries. But there are signs to suggest that governments and militaries are wizening up to this possible reality as seen in the example of the British government recently committing funds to developing counter-drone technologies and Chinese manufacturer DJI implementing tracker technology. While the projects in question are largely focused on countering possible terroristic or criminal applications of drones, they show that debates on drones within Europe and the US possibly shouldn’t be restricted to drones as a tool, but also include the perspective of drones as a risk. 

Most pressingly, the above-named current debates about transparency, technological advancement and ethics are far from theoretical. As seen in a recent report by the NGO Drone Wars UK on Israel’s drone program, drones have found common application in Israeli operations in Gaza in part because they are seen as more humane due to increased precision of technologies and lessened risk to human life. Arduously collected reports and data from the ground seem to suggest, however, that the consequences equal if not surpass the gravity of conventional armaments. In fact, the report concludes that the harm on civilian populations caused by drones is substantial and, moreover, the constant threat posed by this technology causes immense psychological suffering to all inhabitants of the affected region. 

Drones seem to be here to stay, but how they are used hasn’t been set in stone. As the example of the post-conflict use of drones in Rwanda shows, the technology can also serve purely humanitarian causes in the right environment, though concerns about risks and viability might remain. As it stands, the most pressing debates that emerge from these analyses of the last months are: 

• How will the applications and the viability of military drones change if and when the global geopolitical setup changes? 
• How can transparency in the application of combat drones be increased to facilitate a viable public debate on the issue? 
• How will the current course of technical development, particularly machine autonomy, affect the humanitarian effects of military drones? 

• What steps can we take today to avoid facing insurmountable dangers posed by drone technology in the future? 

The debate around drone technology is caught between over- and underestimating its significance. To move forward, it is important to observe the many ongoing developments around drones and to not disregard any of the related facettes of the debate – humanitarian, political, legal – in favour of the merely technological. 

In the last couple of weeks, this blog discussed topics such as the checkered history of drone development, ethical and practical concerns in the automation of wars, the open question of how to defend against drones and the difficulties of reporting on drone warfare. All of these topics appear, directly or indirectly, in a recent report on the Israeli military’s use of drones, compiled by the NGO Drone Wars UK. 

While often overshadowed by the foremost drone-power, the USA, Israel has been among the more active participants in drone warfare in recent years. As mentioned in the brief summary of the history of drones, the IDF was interested in the offensive uses of UAVs early on, making use of them in war as early as the 1970s. While it would be another 30 years before the use of combat drones became commonplace, Israel now operates a successful domestic drone development program and belongs to the top exporters of drone technology. And yet drones occupy an uncanny space between secrecy and publicity in Israel. On the one hand, official sources on drones developed in and exported by Israel are scarce, particularly armed drones. This mirrors the history of the US drone development and deployment program, which was accompanied by denials and non-confirmations until very recently. On the other hand, the report shows that proponents often stress the role of drones in making military missions less risky for operators and bystanders alike. The Israeli military argues, that increased precision facilitated by new technologies will lead to the reduction of collateral damage, framing drones as a means of making combat more humane. 

Drone Wars UK’s report draws these claims into question. In the absence of official sources, analyses of the effects of drone warfare in Gaza have to rely largely on observers on the ground, as is often the case in chronicling drone warfare. The paper compiles various reports of human rights organizations and journalists about civilians who have been killed by drone strikes in non-combat scenarios as well as eyewitness reports and studies of the effects of these strikes. One such study, published in the medical journal the Lancet, suggests that drone strikes are among the most common causes for traumatic amputations conducted in Gaza and that the severity of injuries caused by these strikes tends to be significantly higher than of comparable attacks such as other explosive projectiles. Drone Wars UK’s report documents an immense psychological impact that the near-constant threat of drone attacks causes for the inhabitants of Gaza. Overall, the report concludes that any claims about the humanitarian nature of the Israeli use of drones should be disregarded considering the rising number of casualties and the severity of destruction they cause. 

This case study shows how some of the issues surrounding drone warfare can play out in real-world situations. Fluidity of operations and lack of transparency makes the documentation of drone operations in Gaza a perilous and work-intensive task. But it is precisely this data that is required to accurately evaluate the impact a drone program has, both on military and more pressingly humanitarian grounds. As it stands, the Israeli drone war in Gaza is another case of a completely unilateral use of combat UAVs, with both fighters and civilians on the ground essentially unable to defend themselves effectively. As such, avoiding further civilian casualties can only be achieved by changing the way drones are used. And while some might argue that any of the issues pointed out by Drone Wars UK’s analysis could be lessened by refining existing technologies, it seems that more restrained policies and protocols in the employment of drones are just as neccessary to avoid undue human suffering.

Read the full report by Drone Wars UK here: https://dronewars.net/wp-content/uploads/2019/11/DW-FracturedLives-WEB.pdf 

Rwanda is a country that often remains defined by its conflict-filled past, most notably the devastating genocide that took place there in 1994. But over the last couple of years, the East African republic has become one of the continent’s foremost drone-users. Drones in Rwanda aren’t primarily used for military or intelligence purposes, however. Rather, drones are filling vital functions in a country, in which infrastructure and networks of aid can often be deficient and underfunded. 

In 2015, the construction of a dedicated civilian drone airport in Rwanda was announced and since October 2016, the US startup Zipline has started operations within the country, delivering vital medical goods to remote areas. The company was founded in Silicon Valley in 2014, securing a contract with the Rwandan government shortly after. The delivery drones, visually resembling small airplanes, enjoy the same regulatory status as government flights within the country and Rwandan regulators so far have taken care to facilitate an environment that allows companies like Zipline to unfold their potential. The use of drones that operate beyond the controller’s line of sight is restricted to military uses in many countries, with Rwanda’s delivery drones being a rare exception. As a result, medical deliveries from one part of the country to another can now be completed in 30 minutes where they used to take hours. 

Many roads in Rwanda are rough and vulnerable to flooding and other damages, making deliveries by car time-consuming and less than secure. Medical infrastructure, similarly, is lacking funding, especially in the rural parts of the country, despite government efforts to improve the situation. Medical workers are scarce, and resources limited, making the time and labour savings offered by drones potentially lifesaving. 

The implementation of this drone delivery system came as part of an ambitious modernization effort in Rwanda. The country has since introduced a regulatory framework for personal use of drones and will be hosting a forum about the non-military uses of drones in February 2020. Optimistic voices see Rwanda as a hopeful case study of the peaceful uses of drone technology, that could be implemented around the world if funding is committed and regulators follow suit. On the other hand, some fear that drones are distracting from the lack of basic resources they try to alleviate. The problems that are being solved by Zipline in Rwanda could also be solved by improved planning and more investment in public health, it is said. While the company claims that a delivery by drone costs about the same as a conventional delivery by motorcycle, the exact costs of the project haven’t been made public yet and the operations has some markings of a prestige project. 

Other countries in the region, such as Kenya, furthermore, have been reluctant in adopting similar solutions, due to increased prevalence of terrorist groups within their borders, which might be eager to use UAVs for violent purposes, a risk described in an earlier article. 

The positive results of Rwanda’s drone deliveries can’t be doubted, but how universally this model can be applied remains doubtful. Regulation, again, is a hotbed issue with promises of a futuristic hyper-efficient infrastructure on the one side and the constant threat of abuse by criminal or terrorist groups on the other. What has become clear, however, is that the debate on the role of large-scale drones extends past its military uses, for good or ill. 

A recent post here outlined the difficulties in defending oneself reliably against attacks that make use of drones. While different technological means are being tested, including projectiles, jammers and anti-drone drones, experts concur that so far, the most reliable way to take out a drone is to take out the person controlling it. 

Chinese consumer-grade drone manufacturer DJI, recently unveiled a new app that would make this much easier in the case of improper drone use. The app will use wifi protocols enabling anyone in range to access information about a drone’s course and location, as well as the whereabouts of its controller on their phone. A DJI spokesperson affirmed that the required drone ID functionality will be added to a large number of drones already in circulation through software updates and that the company is coordinating with regulators in implementing the new technology. This development follows the implementation of new regulations making drone IDs mandatory in the US and other countries. 

While this capability is widely seen as a step in the right direction, it is far from a solution to the potential threat posed by weaponized UAVs. The BBC found that experts remain critical about the usefulness of this app in preventing terrorist attack, for example, as they argue that anyone with “real bad intentions” will find ways of circumventing the new system. The implementation is further limited by compatibility issues with older drones and some smartphones. 

So rather than solving the problem of the misuse of drones, this development rather exemplifies how this issue is being tackled on the intersection of regulation, technology and actual physical case-to-case enforcement. And how, regardless, a lot of uncertainty remains. 

The discussion about drones in the English-speaking world, and in the Global North in general, has been largely focussed on when and how to utilise them, as well as some more general questions of principle and practicality. Drone strikes, up until now, have been something that “we” do to “them”, something that happens far away from the towns and country sides of Europe or North America. But the history of every effective military technology is one of proliferation and drones are no exception to this. As drone technology becomes ever more ubiquitous and more refined, one question that has to be tackled more and more is: How can one defend themselves against hostile drones?

This question is not entirely new of course. Those targeted by military UAVs have been struggling to counter the technology from day one and since 2017, there have been cases where even the US military found itself on the receiving end of drone attacks, when insurgents used modified consumer-grade drones to deliver explosives to targets. The fears of ‘drone terrorism’ have increased since then. In a 2015 study, while seeing terrorist attacks via drone as a ‘very low probability event’ (on page 673), Tung Yin, professor at the Lewis & Clark Law School, recommends taking legal and technological precautions to avoid such situations, such as requiring transponders in commercially available drones and a registry of UAVs. But these measures largely seek to minimize the risk of commercial drones being used for terrorism, rather than enabling responses in the case that it happens.

The British military recently committed 2 million pounds to various drone-defence projects, which seek to employ a variety of technologies to counter potential drone attacks. From jamming to blunt force projectiles, the best way to effectively counter a hostile drone remains an open question. This problem is exacerbated by the ever-changing forms that UAVs take. Particularly the ongoing miniaturisation that might bring about ‘drone swarms’ in the future makes simply shooting down drones less and less practicable. On the reverse side, questions can be raised as to how war would change between two sides that utilise large-scale combat UAVs, the way the US currently does.

Drone wars, as we know them, came about from a very specific geopolitical context, but as it appears that military drones are here to stay and warfare with combat drones on both sides will have deep implications for the nature of war. Who fights, who is a target, where (and if) civilians can be safe and how to gather and divulge information about the fighting will all be affected by a shift towards remote warfare. And how armed forces are able to counter drones will most likely be a determining factor in all of these questions. But as it stands, the question how to defend against drones has many answers, few of which are anywhere close to definitive.

See also:

Yin T. (2015) ‚Game of Drones: Defending Against Drone Terrorism‘, in Texas A & M Law Review, Vol. 2, No. 4, pp. 635-673.

War reporting has been a near-constant feature of armed conflicts from the early 20^th century onward. This journalistic coverage of war has served a number of uses over the course of its existence: From propaganda for the purposes of recruitment and fostering political support for the war effort, to informing a civilian populace of potential impending risks and even critically examining the conduct of different involved parties. This practice has likewise undergone many permutations throughout the years. Where newspapers used to print grainy posed portraits of soldiers and armaments a safe distance from the frontlines, we now have live satellite footage filmed directly from fighter planes or reporters embedded with troops on the ground. But combat drones don’t seem to enjoy the same media presence conventional forces do. Footage recorded by surveillance drones is sometimes publicized, as in the case of the aftermath of the recent US operation against the leader of ISIS, Abu Bakr al-Baghdadi, but actual footage of combat drones in action remains scarce. 

This can in part be explained by the semi-official status drones held for a long time. As described in an earlier post, armed drones came to be not only as weapons of war, but equally if not more so as means of counterinsurgency and counterterrorism. Accordingly, within the context of the so-called War on Terror, a large number of lethal drone strikes are committed outside of active warzones, with Pakistan being the most affected countries. It was not, however until early 2012 that the US government officially admitted to conducting a drone program in some areas of the country. According to the Bureau of Investigative Journalism, more than 300 drone strikes, with a potential body count in the thousands, were conducted by the United States in Pakistan between 2004 and January 2012. Even today, the US government remains largely tight-lipped about drone operations in Pakistan. Donald Trump recently cancelled a short-lived executive order, enacted in 2016, that obligated the US government to file an annual report on casualties of US drone strikes outside of “areas of active hostilities”. 

Without a clear limitation of use and with legal and ethical issues still being discussed at home and abroad, it stands to reason that official sources aren’t as eager to publicize drone strikes as they were bombings during the Gulf War. 

This, however, complicates the work done by journalists who track the US’ ongoing drone war significantly more difficult. The above-mentioned Bureau of Investigative Journalism maintains one of the most comprehensive databases of US drone strikes. The organization details the difficulties of chronicling a phenomenon, which spans many years and countries and on which official sources are scarce. In this task, a variety of sources is essential. In warzones there are oftentimes official figures on executed operations and sometimes confirmed casualties. Furthermore, a large number of local and international media sources and in some cases even communiqués by targeted groups are analyzed. A common problem with many of these sources is, however, unreliability. Government sources usually don’t include strikes committed by Special Forces or the CIA and the Bureau reports cases in which the Yemeni government took credit for airstrikes, which it likely didn’t have the capacity to execute. Some local media sources have shown bias in the reporting of drone- and air strikes and even between reliable sources oftentimes discrepancies in reporting can arise. 

The Bureau of Investigative Journalism is not the only organization struggling to document the ongoing irregular war that is waged using drones, but its effort signals a serious problem in the journalistic reporting of asymmetrical wars. The goal of journalism, to foster a well-informed public, which can in turn participate in determining public policy seems to be somewhat out of reach, if the obstacles to find reliable information on a government’s combat operations become bigger and bigger. Drone strikes aren’t invisible, thanks to the work of journalists around the globe, but they seem to exist in a space marked by near-constant uncertainty and large degrees of deniability. 

With every new technology the implicit question about their wider implications arises. New innovations can fundamentally change human behaviour, social conventions and cultural norms. And the question of implications might be all the more prescient in the case of technologies that find use in war. Drones are no exceptions to this. Since the start of their modern use, questions about ethics, changes in military strategy and their effect on global security have been widely discussed. Optimists argue that the use of drones can reduce the net suffering caused by war. Mission that traditionally had to be conducted by human pilots at great risk to their own physical and mental health can now be pulled off by a team that is situated at a safe distance, both practically and symbolically. Critical voices point out that it is exactly this distance that makes drones all the more dangerous. With the, still very real, mostly human, targets further abstracted through visuals on a screen that shows a far distant space, killing might be made easier and the inherent human hesitation to pull the trigger and end a life might be lessened, it is argued. And is it not just this hesitation that often spares a bystander in the middle of the chaos of war? 

These ethical considerations might be about to become more pressing than ever, as artificial intelligence and machine learning start to seep into more and more aspects of modern life, including war. General Atomics Aeronautical Systems, a common supplier of the US military, has recently secured a contract for a new project, the Agile Condor capability, to be tested for use in military drones. The technology aims to facilitate the autonomous identification of targets of interest by a surveillance drone using machine learning, without the need for human input. While in its current state the technology appears to be only intended for use in reconnaissance and targets are supposed to be confirmed by human operators, the implications weigh heavy. Taking a step toward removing the human element from the selection of targets for possibly lethal strikes, or, as some fear, moving towards fully autonomous combat drones, that identify their own strike targets, poses fundamental questions about the ethics at work. 

Artificial intelligence does bear the possibility of accuracy that surpasses human capabilities. How many innocent people have lost their lives in war due to nothing but human error? But those that might be killed due to an error committed by a machine might weigh even heavier, some might argue. A self-expanding algorithm can’t possibly have a conscience or sense for human life, the way that a human pilot, even one controlling a drone from a different country, can have. Reducing decisions over matters of life and death to mere cold calculations, matters of strategy, even more than they are now, not only holds the risk of inhuman choices, but also of normalising them. But while an artificial intelligence won’t ever feel compassion it is also without anger or hatred, impulses for, some might say, some of the worst transgressions ever committed in warzones. 

Automated killing is still, for the time being, more science fiction than reality. But the technological capabilities to make it possible are being worked on and there is little doubt that it will be a possibility sooner rather than later. Questions about whether and how these technologies can and should be used have to be posed and discussed now, not only when the first autonomous combat drone takes flight. 

There are few topics in the debate around military conflict in the 21^st century, that are more current than that of military drones (or unmanned aerial vehicles, UAVs). Still, to not only reach a more grounded understanding of today’s events, but also to be able to think about possible future developments, looking at history is essential. While the use of drones is just one step in the larger continual change that methods of warfare go through, this historic overview will be as brief and focussed as possible without being reductive. This overview will also be focussed on the USA to a large degree, as the US military has been at the forefront of drone development, both historically and today.

Warfare during the 20^th century was marked by a growing importance of aerial combat. From early military aircrafts in WWI to the Gulf War and Kosovo War in the 90s, which could be argued to have been decided in no small part by air superiority, a clear trajectory becomes obvious. And indeed, interest in unmanned, remote-controlled aircraft emerged as early as 1911, when US inventor Elmer Sperry first considered radio-controlled flight as an option. Interest in the topic intensified during wartime and similar ideas started being developed by the British military as the country struggled to wrest aerial control from Germany across the vast frontlines of WWI. Early experimentation on remote controlled flight focussed largely on one-time-use vehicles rigged with explosives or aerial torpedoes, as the radio and flight technology of the time limited the control that could be exercised over a flying object from a remote position. Weapon tests for these early UAVs were conducted towards the end of WWI, but neither the British nor the US military ever made use of any of the early models in combat during that war. It was not until 1924 that the first successful complete flight, including take-off and landing, was executed remotely, using radio technology.

The interwar years also marked the historic moment in which, through improved flight technology and increased destructive capabilities, aircraft started to rival, if not surpass, the military significance of warships. As a result, remotely controlled plans were used by the US navy as targets in combat exercises, in order to prepare for the eventualities of future wars. With the start of WWII, interest in unmanned military flight saw a resurgence. The US Airforce tried remotely steering modified, otherwise obsolete B17 bombers, rigged with explosives into strategic targets. The planes’ take-offs had to still be executed by an on-board crew, who would then parachute from the plane which would continue to be steered from a distance. These experiments, dubbed Operation Aphrodite, proved largely unreliable and resulted in few successful attacks. In 1952, the US Navy again attempted to use obsolete fighter planes as drones to attack strategic targets in North Korea, but again the success rate was low and the project was dropped. With limited use in combat, drones mainly continued to serve as targets for combat exercises until the 1960s.

The changing geopolitical context of the Cold War, however, led to the emergence of a new type of UAV: Reconnaissance drones. While experimentation with combat uses for UAVs continued, the US military saw an increased need for reconnaissance flights and airborne intelligence operations both in and outside of wartime, in order to keep tabs on troop movements, stationed weapons and military developments in the Soviet Union and other enemy nations. Starting in the early 1960s, efforts to modify target drones for surveillance purposes were started and picked up after the Cuban Missile Crisis of 1962 proved the continued potential need for risky recon missions. UAVs were first used for intelligence gathering in 1964, when unmanned drones were used to scout the coast of mainland China, producing usable photographs of military targets, before being successfully recovered in the South China Sea.

US use of UAVs for reconnaissance exploded in the years to come, with thousands of missions over the course of the Vietnam War between 1964 and 1975. A mid-air retrieval system using manned helicopters was implemented. The great success in procuring strategic photography and helping to coordinate attacks during the Vietnam War led to renewed efforts to develop armed UAVs. The ‘Lightning Bug’ drone, that had been used for the vast majority of unmanned missions during the war was used as a basis for this renewed experimentation. The resulting unmanned combat air vehicle was first used by Israeli forces during the Yom Kippur war in 1973 and the Lebanon war in 1982, but despite some successes was found to still be less effective than conventional bomber technology and missile systems by the US military.

Historian Christopher J. Fuller pinpoints the origin of the United States’ modern combat drone program to the counter-terrorism policy of the Reagan Era. Leading figures of Reagan’s intelligence team were looking for ways to perform limited targeted strikes against terrorist groups and saw UAVs as a possibility. Prototypes of drones with propelled rockets are said to have been commissioned by the CIA shortly before the Iran-Contra scandal threw the agency into disarray in 1985. Development continued in the following years with some setbacks, due to prohibitive costs, but ultimately resulted in the development of the RQ-1 Predator drone in 1995, which would become a widely used UAV by both the CIA and the US Airforce first merely for reconnaissance, but later also for lethal strikes. The shock of the 9/11 World Trade Centre attack virtually obliterated institutional resistance to implementing lethal drone strikes, with George W. Bush signing an authorization for the employment of an armed Predator drone by the CIA mere six days after the attack.

The post-9/11 turmoil changed the US counter-terrorism policy profoundly from its Cold War paradigm and facilitated the employment of more militant means against Al Qaeda and the Taliban. The first Predator drone strike was conducted in November of 2001 in Afghanistan. The CIA’s Counter Terrorism Centre conducted its first lethal drone strike in 2002 against a car carrying Al Qaeda fighters in Yemen. The UK military started using Predator drones shortly after in 2004. June 2004 marked the first US drone strike in Pakistan, a notable expansion of drone warfare. The number of total confirmed drone strikes, the vast majority conducted by the US, has grown continually since, reaching a peak during the ongoing presidency of Donal Trump. According to the Bureau of Investigative Journalism, a minimum of 6,786 drone strikes have been committed by US forces since 2004 in Afghanistan, Pakistan, Yemen and Somalia, with a total death toll between 8,500 and 12,000.

This tendency shows, that UAVs, as weapons of war, are the goal of a century-long aspiration by some military minds, but that the use of armed drones by US forces as we see it today is based in the particular environment created by the post-2001 War on Terror. This also suggests, that the use of combat drones, both within and outside of war zones is not likely to end any time soon as the USA’s counter-terrorism efforts show little sign of subsiding and few can challenge the US drone war on the global political stage. If we accept this assessment, this makes questions of proliferation, sovereignty, accountability and reliability that drone attacks pose all the more pressing in the present.

Sources:

Keane, J.F. & Carr, S.S. (2013). A Brief History of Early Unmanned Aircraft, John Hopkins APL Technical Digest, Vol. 32, No. 3, pp. 558-571.

Fuller, C.J. (2015). The Eagle Comes Home to Roost: The Historical Origins of the CIA s Lethal Drone Program, Intelligence and National Security, Vol. 30, No. 6, pp. 769-792.

Rise of the Reapers: A Brief History of Drones, by Chris Cole on Drone Wars: https://dronewars.net/2014/10/06/rise-of-the-reapers-a-brief-history-of-drones/

Drone Warfare, data by the Bureau of Investigative Journalism: https://www.thebureauinvestigates.com/projects/drone-war