in january 2016, a small, white Cessna aircraft began to hover over the city of Baltimore.1 Although the plane was indistinguishable from hundreds of ordinary private aircraft that are seen high above many US cities, this one had a very different purpose. While it was not acknowledged at the time, this Cessna flight was part of an effort toward placing a major US city under the kind of persistent, wide-area surveillance that had been pioneered by the US military in conflict zones around the world. Like many drones, this manned aircraft was equipped with sophisticated, real-time video and communications capabilities that would allow it to capture grainy, black and white images and convey them to analysts on the ground. These analysts, tucked behind a bank of computer monitors in an unremarkable office building in downtown Baltimore, could produce new evidence, and sometimes photos, of crimes reported to police. And the cameras were capturing a lot: when the Cessna was aloft for five to six hours a day, its wide-angle lenses could capture 32 square miles of ground per second.2 At a typical height of 8,000 feet, it could capture a significant portion of the city each day provided it went on one or more flights. The analysts could see cars moving and people walking around the city in real time, but the images were not yet detailed enough to see features of people’s faces or their clothing. For the Baltimore Police Department, which had sponsored these flights as test runs, this was enough. The images coming from the Cessna allowed them to track the movement of cars from hit-and-run incidents or to find which buildings people fled into after crimes had taken place. All of these were clues that allowed the police to hunt down suspects, even days later, who might otherwise have escaped capture. Wide-area surveillance seemed to offer a vision of the future for law enforcement and gave the Baltimore police department some hope of stemming the tide of violent crime that had been afflicting the city.
Like most developments associated with drones, the origins of these surveillance flights lie with the US military’s investment in research and development of unmanned aerial technology. After the invasion of Iraq in 2003, the US military faced a vicious, multi-faceted insurgency that threatened to destabilize the new government. Perhaps the greatest threat for US troops was IEDs: the Iraqi insurgents proved resourceful in developing crude explosives to target US convoys and personnel to devastating effect. Roadside bombs were often buried along major routes taken by convoys or concealed under piles of trash and the carcasses of dead animals. Hiding in the civilian population, insurgents detonated the bombs by remote trigger as US forces moved through towns and villages. As the insurgency wore on, the IEDs became more sophisticated and deadly, in part because Iran was supplying technology to aid the bomb makers.3 As the death toll grew, the United States began to uncover bomb factories across Iraq where insurgents tried to make these explosives on a near-industrial scale. After its success in Iraq, the tactic spread to other theaters, including Afghanistan, where even more US troops found themselves thwarted by crude improvised explosives. One Pentagon estimate in 2013 suggested that between half and two-thirds of all US casualties in Iraq and Afghanistan were due to IEDs.4
Facing such a serious threat, the Pentagon began to invest billions of dollars into counter-IED technologies that either prevented attacks or shielded US troops from their consequences. Among the many contractors that the Pentagon worked with was a private company called Persistent Surveillance Systems (PSS), which was owned and operated by Ross McNutt, an MIT-trained engineer who had worked for the Air Force. He developed a wide-area surveillance system that would allow the Pentagon to record real-time video over an entire city. One crucial aspect of the system that he developed, initially called Angelfire, was that it could rewind through hours of video and identify who placed an IED once it had been located. This meant that US intelligence analysts could rewind to find the insurgent who placed it, no matter how long ago it was placed. The movement of the bomb placer could then be traced over time back to safe houses or the IED factories that had sprung up all over the country. Analysts could then figure out who this person spoke to and what places he visited in order to draw a picture of the social network lying beneath the insurgent cell. This capability was ideally designed for rolling up the networks of insurgents who had made the emplacement of IEDs so skillful and deadly. The program, pitched by McNutt as “Google Earth with TiVo capability,” was purchased by the Pentagon and transferred to Los Alamos National Laboratory in New Mexico, where it was combined with other prototypes and put into service.5 Like many military surveillance aircraft, it was also equipped with night vision and EO sensors, as well as signals intelligence and geolocation sensors, to provide a full-spectrum stream of data on the insurgents placing bombs in the way of US troops.6
The surveillance program for IED detection reduced combat deaths for US personnel, but it did not last long.7 Once US personnel began to leave Iraq at the end of formal combat operations in 2011, the program fell into disuse. Like many former military contractors, McNutt began to explore uses for surveillance inside the United States itself. He ultimately settled on offering Baltimore a pilot program, in part due to the severity of the violent crime problem that the city faced and in part because he had personal contacts there.8 Baltimore’s rising crime rate made it an attractive potential test site, and PSS was convinced that they could lower crime rates by 20%–30% if their system was implemented on a full-time basis.9 Private philanthropists donated the funds to the Baltimore Community Foundation, a non-profit involved in civic work in the city, which in turn activated the contract with PSS and began feeding the data to the police.10 Their contribution was important because it was an expensive operation, costing between $1,500 and $2,000 per hour of flight.11
The secret to making the pilot program work in Baltimore was to keep the program as quiet as possible. The contract with PSS was never publicly announced or discussed at government hearings. Baltimore’s mayor, Stephanie Rawlings-Blake, admitted that she never heard of the program until the story was broken by Bloomberg news in August 2016.12 The Baltimore police department retitled the program, in anodyne terms, the “Community Support Program” and its offices were located in a nondescript building in the center of the city. In some cases, police officers who received tips or even images from the Cessna did not know where the information was coming from. By some measures, the surveillance worked: police officers were able to track individuals and vehicles, provide briefings with leads to investigators, and increase the percentage of crimes “cleared” by the Baltimore police department. The independent Police Foundation reviewed the program and concluded that it was effective, but conceded that it was not well explained to the population at large.13 The Baltimore Community Support Program even argued that the cameras were capable of deterring crime, although measuring whether crime has not happened for a particular reason like the presence of a camera is difficult.14 The Police Foundation also concluded that its success could be replicated elsewhere.15
Despite its apparent effectiveness, the program remained highly controversial. The presence of a Cessna aircraft collecting video on a near-constant basis over a city violated what many saw was a reasonable expectation of privacy. The PSS program captured video of thousands of people who were not charged with a crime and who expected (incorrectly) that many of their actions at that point were entirely private. In particular, it violated an expectation of privacy in non-public places. While many people have accepted that they are now being filmed by closed-circuit TV cameras (CCTV) in shops and public places, they do not accept being filmed in their backyards or other private venues. As critics have noted, people engage in activities in presumably private venues that differ significantly from their behavior in public, thus rendering surveillance of private and semi-private locales more revealing and perhaps invasive.16 Persistent surveillance also reveals people going to specific locales—such as mosques, doctor’s offices, AA meetings, gun shows, abortion clinics, and so on—which can reveal their religious and political preferences. Some of this information could be personally damaging if turned over to an employer or an insurance company.17 The program was supposed to hold video footage for only forty-five days, but in practice the Baltimore police broke this promise and kept the video for much longer, with no notice.18
Another potential danger is that persistent surveillance could reinforce the biases already existing among law enforcement organizations, rather than correct them. When PSS was revealed, it immediately raised questions about whether it had been used disproportionately over African American neighborhoods. Defenders of the system argued that the cameras did not target any one neighborhood in particular and that its coverage was balanced between the white and black areas of the city, as well as wealthy and poor parts.19 McNutt argued that the cameras could not yield facial or skin details to make such racial profiling relevant; however, the PSS data is used in tandem with other CCTV cameras and could, at least in theory, be combined for racial profiling.20 The danger that surveillance technology could be used disproportionately over “suspect” areas makes it particularly alarming from the perspective of fairness and equality.
The PSS has also raised new questions about freedom of protest and government accountability. With no public notice and operating entirely in secret, the PSS took as many as one million photos of daily life in the city of Baltimore.21 Most of those photos were of routine life, but the American Civil Liberties Union (ACLU) noted that McNutt and his employees were alert to the movements of protesters, such as those coming from Black Lives Matters groups, and monitored the city for signs of protest after some controversial verdicts in criminal trials.22 Watching for signs of protests goes far beyond delivering videos of existing crimes and poses a danger to the right of peaceful assembly guaranteed under the First Amendment. Much of this surveillance was also done without any evidence of democratic consent. There were no city council meetings to approve what was essentially a privately funded venture.23 The Baltimore Police Department tried somewhat evasively to defend the PSS program as being part of CitiWatch, a network of stationary CCTV cameras already in use. But the scope and mobility of the PSS camera—in effect, the fact that it can see anything anywhere—suggests instead that this technology was different, and perhaps more dangerous, than a fixed camera system.24 Entirely in secret, and without a single public hearing, the Baltimore Police Department had put the entire city under watch but left no one officially accountable for misuse of the imagery flowing from their cameras.
All of this was done using a Cessna aircraft operated by a pilot. In many respects, the surveillance was therefore limited: the plane could remain in the air for only six hours at a time, and the practicalities of human endurance meant that night flights and flights in bad weather were often ruled out. But when surveillance capabilities are combined with the endurance and adaptability of unmanned aircraft, many of these natural limits fall away. In the future, it is likely that cities will attempt this kind of “eye in the sky” level of surveillance with drones, which are cheaper and can operate for longer periods of time than their manned counterparts. There is already evidence that drone manufacturers like DJI and AeroVironment are specifically targeting police forces that cannot afford their own small planes and helicopters and offering them specialized law enforcement drones for surveillance and other tasks.25 While much of the debate over the rise of drones in law enforcement has revolved around issues of privacy, there are two other equally important issues at stake. The first is goal displacement. Does the availability of surveillance cameras and videos combined with drones make users of all kinds—law enforcement, dissidents, journalists, and even governments—“throw the dice” and engage in activities than they otherwise might otherwise have been reluctant to undertake? The second issue concerns the risks of employing this technology for political life more generally. If governments are equipped with surveillance drones, will this state of affairs erode the right to privacy? And will it lead to greater repression of dissent, and a darker future ahead, for those living under authoritarian governments?
In US skies today, drones are used for a growing variety of purposes from surveying real estate to tracking wildfires and delivering pizzas. The long-heralded commercial boom of drones—in which Amazon will send drones for “last mile” package delivery and most commercial goods shipments from DHL, FedEx, and others are operated by drones—is not yet here but will probably get underway within the next decade. The current pilot programs for drone delivery are generally restricted over cities or other high population areas due to FAA regulations and public nervousness about what would happen if they came down unexpectedly.26 Still, at present, the US government estimates that 110,000 commercial drones are in the skies and that up to 450,000 may be operating by 2022.27 Although progress has generally been slower than expected so far, the FAA is under intense pressure from lawmakers to open US airspace to commercial drones and to enable retailers like Amazon to realize the cost savings that “last mile” drone deliveries will yield. In 2018, the FAA authorized a new series of pilot tests for drone deliveries and proposed new, as yet unapproved, regulations governing how and where commercial drones could move through US airspace.28
Yet while the skies will soon hum with more and more drones, that does not necessarily mean that surveillance of the US population will also increase in equal measure. Most domestic or civilian drone use has little, if anything, to do with surveillance, and rather remains commercial or scientific in nature. One estimate of non-violent drone usage from 2009–2015 found that surveillance constituted only 9% of the global usage of drones, far less than the commercial or scientific deployment of the technology.29 Even a police or other law enforcement drone will not necessarily always be conducting surveillance, or even in most cases. The drone that is hovering over someone’s backyard is far more likely to be a hobbyist drone or a drone mapping local real estate than a surveillance drone.
What has changed is the number of actors capable of launching a drone with surveillance capability. The spread of drone technology has effectively democratized the US airspace and opened up opportunities to fly independently to those who would never otherwise have it. Until recently, the ability to fly one’s own aircraft was a privilege given only to those who could afford it. Only the rich with private planes and powerful companies could field their own aircraft; almost everyone else who wished to fly was forced to take commercial aircraft. This meant that only a select few—predominantly government agencies, but also some private entities like universities and private companies—could get “eyes in the sky” and spy on activities below. As a result, aerial surveillance was largely limited to the government, law enforcement, and select businesses such as high-end real estate. That is not the case today. Now virtually anyone can get “eyes in the sky” and conduct some form of crude aerial surveillance. Everyone from real estate companies to paparazzi to private investigators can now get a bird’s-eye view of their target and use drones to take video and film from the air. So, while the drone hovering over the backyard is probably not a surveillance drone, it will be harder to tell who it belongs to if it actually is.
At the same time, the use of drones for surveillance in domestic airspace is increasing. Inside and outside the United States, law enforcement organizations are now seeking drones for a variety of uses, one of which is getting an aerial overview of their territory. In April 2017, a study by the Center for the Study of the Drone at Bard College estimated that 347 state and local police, fire, and emergency services have purchased drones.30 Most drones are held by sheriff’s offices, police agencies, and fire departments, though other first responder organizations have also bought drones.31 At this point, the kind of sweeping surveillance conducted by military grade technology, as seen in Baltimore, is relatively rare. The vast majority of drones purchased by law enforcement organizations are relatively small quadcopters, which can remain in the air for only a short period of time and have a limited capability for capturing images and video. Unarmed, camera-equipped drone models such as the Skyseer, the Skyranger, T-Hawk, WASP III, Shadowhawk, and a number of DJI models have all been used by domestic law enforcement agencies.32 Law enforcement agencies in more than a dozen US states have sought approval from the FAA to fly drones. A number of large cities, including New York and Los Angeles, have also purchased drones for their police forces. As one New York Police Department spokesperson remarked, drones “aren’t that exotic anymore.”33 Private companies are also eager to capitalize on the demand for drones among law enforcement. For example, DJI has partnered with a company called Axon to sell drones directly to police departments and other public safety agencies.
Many of the missions envisaged for police drones have little to do with constant surveillance but rather involve their use for time-limited, specific tasks. For example, some small police forces have purchased drones to search rural areas for endangered or missing children, to survey wildfires and other natural disasters, to hover over active crime scenes (for example, hostage situations), and to provide intelligence to officers on the ground. Drones can reach remote crime scenes relatively quickly and alert emergency services perhaps before the police even arrive; in other cases, they might also be able to look into windows of buildings to find active shooters and other threats.34 Drones can also be used to watch large gatherings of people—for example, music festivals and events like the Superbowl—in order to guard against terrorist attacks.35 With some adaptation, they can also be deployed as mobile phone base stations and intercept calls before they go through.36
Some law enforcement uses go beyond the limits of a simple camera-equipped drone in the sky. For example, Customs and Border Protection (CBP), an agency which operates under the auspices of the DHS, has been deploying Predator drones and their maritime equivalent, Guardian drones, for border surveillance since 2005. The CBP drones are also sometimes flown for the benefit of other government agencies including the Drug Enforcement Agency (DEA).37 Operating out of Texas, Arizona, and North Dakota, the CBP’s fleet of ten drones has focused on routine patrols over the southern border, especially for surveillance and interdiction of drug shipments, but it has also been used for emergency and disaster response. In 2008, the CBP began to use drones to monitor the northern border with Canada.38 By 2010, the United States had expanded drone patrols to watch the entire southern border with Mexico.39 The CBP’s Predator and Reaper drones are capable of filming video and detecting signals on the ground, and at least one Reaper was equipped with wide-area surveillance video capability that could capture an area approximately 3.7 miles wide.40 Although the CBP planned to expand their drone fleet, they ultimately ran into criticism over the effectiveness and cost of the program. The CBP’s drones were involved in only 0.5% of the total apprehensions and drug arrests conducted by the agency between 2013 and 2016, at a cost of more than $60 million a year.41 The drones were also expensive per flight hour. In 2013, an internal investigation found that they actually cost $12,255 per hour—not $2,468 as had been estimated by DHS—and even this excluded the costs of pilots, equipment, and overhead.42 For this reason, the inspector general of DHS concluded that expansion of the medium-altitude drone program was unwise. There have been also concerns over data protection and privacy violations from CBP’s drone surveillance, especially as the drones can fly as far as 60 miles inland from the southern border and 100 miles from the northern border.43
Under pressure to cut costs, the CBP has also experimented with using smaller drones, such as Shadowhawk drones, to monitor the southern border. But this has turned policing the border into a game of drone versus drone, especially against the Mexican cartels. Since 2010, Mexican cartels have used commercial drones, like the DJI Phantoms, to bring small packages of drugs across the border. In 2012 alone, US law enforcement interdicted 150 different drones carrying cocaine, heroin, and marijuana.44 Drones are also used as spotters to identify safe routes for traffickers to cross the border. The advantages of commercial drones for this practice are clear. Unlike drug runners, drones can tell no tales, and there is little risk to the organization if one is shot down. Drones are particularly cost-efficient now that Mexican cartels have invested in indigenous drone production. Building their own drones also saves the cartels the risks involved in buying commercial models made by other countries, which might have in-built GPS trackers and other revealing technology. The Mexican cartels have also been upping their game in terms of what drones are capable of doing. In 2017, a drug runner was caught with a commercial drone packed with a homemade bomb, the first time that law enforcement saw the use of weaponized drones along the border.45 The United States has been deploying smaller drones as a surveillance substitute for the medium-altitude drones like the Predator to stop these threats along the border, but have been confounded by GPS disruption and spoofing by the more sophisticated Mexican cartels.46
Although this shows how drone use can expand over time, it is worth noting that this sweeping surveillance effort along the United States-Mexico border is beyond the reach of most normal police departments. DHS has given grants to local law enforcement to jumpstart their drone programs, but many grants are not enough to buy some of the sophisticated technology on offer. The result is stark variation, with some well-funded police departments purchasing military-grade drones and others getting by with much less. For example, the Arlington, Texas, police department purchased the Leptron Avenger, a large helicopter drone that can fly and take pictures at up to 12,000 feet.47 Similarly, the Miami Dade police department has been using Honeywell T-Hawks, which can fly at up to 10,000 feet, record images in daytime or with infrared cameras, and zoom to follow suspects.48 Small and rural police departments have been buying commercial drones equipped with thermal scanners and biometric tools not widely available to the public.49 Some of these drones are also equipped with night vision cameras.50 These purchases are driven in part by the need to cover large amounts of ground on a small budget.51 For example, in 2011 the police department in relatively small and rural Mesa County, Colorado, was among the first to win FAA approval to fly Dragonfly drones all over their jurisdiction.
For police departments with the fewest resources, the best available option has been to buy off-the-shelf commercial models like the DJI Phantom and DJI Matrice drones. In 2017 the Center for the Study of the Drone found that most of the drones purchased by law enforcement and other public safety organizations fall into this category.52 Even so, the spread of drones across law enforcement organizations has been wildly uneven. Some major cities have been foregoing drones because of tight budgets, while smaller police departments are investing in them when they have a surplus budget or just a high level of interest among their officers. The result is a patchwork of drone coverage: some counties will have law enforcement drones while their neighboring counties, equivalent in all other ways, do not. It is then hard to know whether a specific police organization will have drone assets that they can call in. Some law enforcement organizations who lack their own drones may be able to borrow drones from private citizens or other government departments to respond to emergencies. Some areas will have no drone coverage at all. It comes down to resources, local initiative, and the degree to which high-ranking law enforcement authorities want a particular city or region—like Baltimore—under close watch. Over time, the United States might become even more unequal regarding aerial surveillance, with citizens in some areas living in full privacy but others under surveillance, either by their own choice or because of the choices of those more powerful than themselves.
Outside the United States, law enforcement organizations have been using drones for an increasing variety of tasks, including monitoring crowds at high-profile events, pursuing criminal inquiries, and even for routine work like stopping speeding motorists. Some European governments have begun to experiment with using drones for more invasive law enforcement tasks, which the US public would not support, such as for issuing speeding tickets.53 For example, the London Metropolitan police plan to use drones to watch for dangerous driving, such as road racing, which may endanger other motorists.54 Police in Kent, England, have begun working with BAE Systems, a British defense manufacturer, to conduct routine monitoring of anti-social motorists, agricultural thieves, and even those who dump trash in protected public areas.55 Similarly, in Ireland, the government is contributing funds to purchase drones to catch those dumping waste illegally, especially in rural areas.56 In France, drones are used to take aerial images of reckless drivers for future prosecution.57 In general, European law enforcement officials have been more willing to experiment with using drones to encourage good social behavior, such as obeying the speed limit and not littering, than simply stopping bad behavior as their US counterparts have.
In the European Union and the United States, the deployment of drones for law enforcement tasks has triggered a renewed debate about the balance between security and privacy. The EU position on this balance is complex and multi-layered: in general, EU regulations impose more legal limits on the use of data by private companies and permit citizens to have damaging stories removed from the internet in the interest of their privacy.58 The European Convention on Human Rights explicitly recognizes the right for “private and family life.”59 At the same time, many EU governments have responded to fears of terrorism by increasing the surveillance power of the state and reducing the privacy protections of their citizens. Drones are one element of this surveillance power, but in terms of raw numbers they have been eclipsed by the growth of CCTV and other more intrusive forms of surveillance, including the bulk collection of communications data. For example, in Britain alone, one 2011 study estimated that there was one CCTV camera for every thirty-two people in the country.60 Similar efforts to ramp up CCTV camera coverage of public transportation hubs and other major tourist spots have been made in France and Germany. In these countries, the CCTV cameras are not always owned by the government, as many private businesses have extensive CCTV networks whose footage can be pulled and used by law enforcement in criminal investigations. In countries where CCTV cameras are more dispersed, such as Italy, other means of surveillance are common. For example, Italy features more bulk collection of communications than many other EU countries, according to a 2008 estimate by the BBC.61 In recent years, most major EU governments have passed new legislation boosting their capacity to spy on their citizens and to share data with EU authorities and sometimes the United States.62 The uneven spread of police drones across the European Union, while important, pales in comparison to the surveillance collection capabilities that these other tools yield for EU governments.
In the United States, the adoption of drones by law enforcement has been challenged at the federal and state level on the grounds of privacy. Although the US Constitution does not expressly offer a right to privacy, the Supreme Court has ruled that various amendments—specifically the Fourth and Fifth but also the Ninth—together yield a right to privacy for US citizens. This right is also enshrined in Anglo-Saxon common law and incorporated within a number of federal and state laws.63 The minimum standard for privacy protects the security of one’s person and property from search and seizure without a warrant. This protection generally extends beyond one’s immediate bodily person to include one’s home, vehicle, or place of business. The interpretation of the right to privacy held by the Supreme Court has traditionally been a physical one, deemed violated only when police directly apprehend people or conduct searches of their homes without a warrant, although some states may interpret this right more broadly.
In applying this standard, the Supreme Court has generally been willing to permit the government to conduct surveillance provided that it does not lead to a direct or illegal search or seizure of a person or property. For example, in Olmstead v. the United States (1928), the Supreme Court upheld the government’s decision to tap the phones of Roy Olmstead, a notorious bootlegger who has been importing and selling alcohol in violation of the government’s laws during Prohibition.64 The Supreme Court found that wiretapping surveillance did not violate Fourth Amendment rights provided that they did not directly search or seize the person or effects of the subject. In a dissent, Judge Louis Brandeis argued that this narrow interpretation of a violation of the rights to privacy did not account for technological change and the capabilities and powers that wiretapping and other new technologies offered the government.65
In fact, Brandeis had recognized the core of this problem—that technology can radically challenge one’s expectations of, and rights to, privacy—in an influential article published three decades earlier. In a scholarly article co-authored with Samuel D. Warren in 1890, Brandeis had offered a reinterpretation of the right to privacy and recast it as the right to be “left alone” from eavesdropping and other intrusions into private spaces.66 Their interpretation was a reaction to the invention of the camera, which enabled a celebrity-obsessed press to photograph and record movements in or near one’s home or place of business. Their goal was ultimately to locate a standard of privacy that would “protect the privacy of the individual from invasion either by the too enterprising press, the photographer, or the possessor of any other modern device for recording or reproducing scenes or sounds.”67 Applied to today’s technology, the “right to be left alone” standard would protect individuals from drones taking photos and videos of them in private spaces, even if the drone never landed on their property or violated the airspace above. But as Olmstead v. the United States and other cases demonstrated, the government and courts were slow to adopt this broader standard and generally deferred to law enforcement in cases concerning surveillance.
In 1967, the Supreme Court finally overturned the Olmstead ruling and offered a new legal test for applying privacy standards in the case of Katz v. the United States. The Court found that the FBI violated the Fourth Amendment rights of alleged bookie Charles Katz by attaching an eavesdropping device to a public telephone that he was using.68 The court found that wiretapping constituted a search and hence was a violation of Fourth Amendment rights if conducted without a warrant. In a famous concurrence, Judge John Marshall Harlan II offered the “reasonable expectation of privacy” test, later known as the “Katz test.” This test requires the subject of surveillance to believe that their activities were private at that moment in time, even if it fact this was not completely true. What this means in practice remains debatable—ultimately, it is hard to know when someone has a reasonable (but still subjective) sense of privacy—but it suggests that privacy can be violated even when no physical search or seizure is conducted. Yet the Supreme Court has acknowledged limits to this standard as well. For example, activities which are clearly in the public view—for example, gardening in one’s backyard—would fail the reasonable expectation of privacy test. In general, the Supreme Court has found that one does not have a reasonable expectation of privacy for behavior which may be seen accidentally by bystanders or others in the immediate vicinity.69 It has also ruled that there is no “reasonable expectation of privacy” in public areas, such as train stations, airports, and open roads. In these environments, one may be openly subject to CCTV or other forms of aerial surveillance, such as drones, without notice or a warrant.
The implications of the Supreme Court’s “reasonable expectation of privacy” standards are important for understanding what role drones may play in extending surveillance. In public areas, the absence of any privacy protections means that law enforcement can fly drones overhead and take photos and videos of individuals without consent. There are few, if any, restrictions on what can be filmed in public areas. In private areas, one has a reasonable expectation of privacy except when criminal activity falls into plain view. In a number of prominent cases, aerial surveillance did produce evidence of criminal activity inside a person’s home. For example, in California v. Ciraolo (1986), the Supreme Court ruled in a five to four decision that police officers acting on an anonymous tip did not need a warrant to fly a plane over Dante Ciraolo’s backyard to search for marijuana plants.70 This case implies that the taking of pictures and videos of the public areas of private residences and places of business during aerial surveillance, conducted either by manned or unmanned aircraft, would not necessarily be prohibited, even if carried out without a warrant. Similarly, in Florida v. Riley (1989), the Supreme Court found that no warrant was needed to fly a helicopter only 400 feet over the property of a marijuana grower’s property. But this again exposed the degree to which new technology is throwing up challenges to existing standards of privacy. In a famous dissent to the Florida v. Riley case, Justice Brennan wrote:
Imagine a helicopter capable of hovering just above an enclosed courtyard or patio without generating any noise, wind or dust at all—and, for good measure, without posing any threat of injury. Suppose the police employed this miraculous tool to discover not only what crops people were growing in their greenhouses, but also what books they were reading and who their dinner guests were. Suppose, finally, that the FAA regulations remained unchanged, so that police were undeniably “where they had a right to be.” Would today’s plurality continue to assent that [t]he right of the people to be secure in their persons, houses, papers and effects, against unreasonable searches and seizures was not infringed by such surveillance?71
As Matthew Feeney from the Cato Institute has argued, this description of the hypothetical device is eerily close to today’s drones.72 It is not unthinkable that police drones could be found hovering in one’s backyard, or just outside a window, collecting incriminating information about a person. Even the collection of that information without a warrant would not stop police from getting a warrant for a full search of the property under existing US law.
A natural reaction to this state of affairs is to insist that police and private drone operators—for example, real estate agencies, private investigators, and so on—would be limited by property rights, especially for private homeowners. A homeowner might assume, for example, that they would have the right to knock a drone out of the sky if it was hovering just over their backyard. But would someone be able to shoot down or disrupt a drone over their own house because it has trespassed on their property? The original interpretation of property rights grounded in common law would presumably permit such an action based on the traditional principle of what is known as cuius est solum, eius est usque ad coelum, which translated from Latin means “to whomever the soil belongs, he owns also the sky.”73 It is this principle that asserts ownership of not only the skies above one’s house but also whatever lies underneath the ground of the property itself. But this principle was tested in the case of United States v. Causby (1946) and was found, in the words of Justice Douglas, to have “no place in the modern world.”74 In this case, Thomas Lee Causby owned a chicken farm outside Greensboro, North Carolina, that happened to be near a military airport. Regular military flights passing 83 feet above the ground flew too close to his home and startled the chickens, causing them to panic and sometimes injure or kill themselves. The Supreme Court found that Causby did not have full ownership of the entire airspace above his property, as the traditional principle would hold. In other words, a homeowner does not have unlimited control of the airspace above their house. However, the Supreme Court held that the overflights over his property did constitute a violation of Causby’s property rights provided that it did substantially interfere with his use of the land and his enjoyment of the property.75 The result of this verdict was to affirm commercial and private flight over private land but also to insist that it must be conducted at a sufficient height not to materially affect the owner of the land itself.
The central question that emerges here is at what height must an aircraft, manned or unmanned, be flown to violate the Causby standard? Some subsequent interpretations of the Causby standard held that there must be some designated navigable airspace—perhaps 400 feet and above—where commercial and private aircraft and drones may freely fly, though the courts have not always upheld this strictly.76 At a minimum, the FAA insists that 400 feet is the minimum safe altitude that aircraft and drones must maintain over congested areas and exercises its own regulatory power on aircraft in that range and above. Below that altitude, there are exceptions made for model and toy aircraft, which can include some drones depending on their size. The FAA permits small aircraft like toy planes and some small drones to fly freely up to 400 feet with a more relaxed set of regulations (such as not requiring pilot’s licenses for an owner). From a surveillance vantage point, this suggests that law enforcement and government drones flying in navigable airspace above 400 feet are in the clear, legally, and even some small drones may be permitted overflights over private property at a lower altitude. All of this suggests that ordinary citizens’ privacy protections against drone incursions are relatively weak.77
In response to this weakness and uncertainty, a number of state and local governments began to pass laws concerning drone usage and violations of privacy. As of 2018, forty-one states had passed laws regulating the use of drones in their territories, and three more had passed resolutions.78 Most have been motivated by the sense that FAA regulations and legal protections for privacy do not go far enough given the advance of drone technology. Yet state and local officials lack the appropriate jurisdiction to regulate the airspace, clearly given to the FAA under federal law.79 Although there are relevant privacy laws at the state and local level, these do not supervene the FAA’s jurisdiction surrounding the use of drones in the airspace. Some states have also passed legislation permitting homeowners to shoot down drones over their properties. At a minimum, permitting drones to be shot down is dangerous because individuals may miss the drones and send ammunition off into the distance to harm others.80 But the FAA has plainly stated that it is a violation of federal law for an ordinary person to shoot down a drone, much in the same way that it is illegal to interfere with a manned aircraft.81 From a legal vantage point, a drone is classified as an aircraft just like a private plane is. In fact, the DHS, among other federal government agencies, is seeking authorization by the FAA to shoot drones out of the sky if a “credible threat” from a drone emerged.82
In the end, this suggests that the opportunities for government or even private surveillance using drones are substantial. Although it is commonly held that individuals have a right to privacy, the Supreme Court has generally sided with the government when it conducts surveillance of a target without a warrant using new technology such as wiretapping devices, GPS trackers, and now drones. The courts have also created a generous interpretation of navigable airspace that suggests surveillance above 400 feet is permissible even over private property. While there are greater protections for privacy below that altitude, it remains unclear what protections, if any, exist for those who wish to stop routine drone incursions. The laws permitting drone overflights are likely to be clarified and possibly expanded once commercial drones from Amazon and other companies take off. At a minimum, the FAA’s default position—that drones should be treated as equivalent to aircraft—restricts the ability of subjects of surveillance to interfere with drones that are hovering over their backyard. This fact, coupled with the absence of restrictions on drones in public spaces, suggests that the drone technology itself will gradually weaken the presumption of privacy currently assumed by many people.
In some cases, this may be a good thing. A world full of drones, and one in which privacy is harder to come by, may be a world in which light falls on injustice and corruption. A number of drone activists have argued that with drones in the hands of more citizens government officials of all kinds, including police officers, will be less likely to exercise their power capriciously. In the United States, a number of press organizations have begun to use drones to cover news stories, though they have run into regulatory and legal barriers imposed by the FAA over the commercial use of the technology.83 Once “drone journalism” flourishes, it is not hard to imagine the technology being used to monitor government or corporate misdeeds, or police abuses. Just as cell phones have been used to record evidence of corruption and abuse, drones equipped with cameras and video equipment may also be put in the service of keeping government honest. For example, in 2017 protesters against the Dakota oil pipelines deployed drones to monitor the progress of the pipeline but also to expose evidence of police using tear gas and water cannons to disperse them.84 Their activities led to a number of proposed state laws declaring the use of drones in this way to be an attack on critical infrastructure.85 Journalists have also used drones to fly over and photograph immigrant detention camps along the southern border of the United States to defy the Trump administration’s attempts to limit scrutiny of the camps by Congress and members of the press.86
Outside the United States, there is also opportunity for drones to be put in the hands of human rights activists to record government repression, violence, and corruption. Deployed correctly, drone activism can shed light on human rights abuses and possibly motivate the world to act. As Andrew Stobo Sniderman and Mark Hanis from the Genocide Intervention Network have argued with respect to atrocities committed by the government of Bashir al-Assad in Syria:
Imagine if we could watch in high definition with a bird’s eye view. A drone would let us count demonstrators, gun barrels and pools of blood. And the evidence could be broadcast for a global audience, including diplomats at the United Nations and prosecutors at the International Criminal Court . . . The better the evidence, the clearer the crimes, the higher likelihood that the world would become as outraged as it should be.87
While it remains unclear whether drone footage of atrocities would be sufficient to motivate states to act, it would at the minimum supply some video evidence that might be usable at the International Criminal Court (ICC). It is possible to imagine that video evidence from drones will be an important complement to other forms of evidence used to hold the perpetrators of human rights abuses to account.
One of the key questions regarding drone activism will be who owns the drones and the pictures and footage coming from them. Some drone activists are hopeful that the United States and other powerful governments with drone assets will be willing to use the pictures and video that they produce to either publicize abuses or provide evidence in court cases against human rights abusers. However, it is not always clear that the US military and others would do so. At a minimum, turning over drone footage would reveal their ability to watch and record events on the ground, which the United States generally tries to protect as secret. But also, admitting that a government has footage of human rights abuse may trigger some international legal obligations to investigate it and may, in some circumstances, yield legal responsibilities to act for the state recording the footage.88 For example, if the United States detected evidence of genocide with a drone overflight, it would be positively obliged to act under international law to prevent the genocide from continuing. Given this possibility, it is likely that human rights activists will find that governments vary in their transparency and willingness to cooperate regarding the detection of human rights abuses with their drones.
For this reason, a number of human rights advocacy organizations have begun to experiment with using drones of their own to collect footage and shame governments into acting. For example, Human Rights Watch used a Sensefly drone in 2017 to fly over waste sites in Lebanon to uncover evidence of the burning of household waste and prove allegations that the government was failing to curb this unhealthy practice.89 Yet human rights organizations using drones will face some of the same problems that governments do when protecting their intelligence sources and methods. Almost all forms of digital or video recording—whether from a cell phone video or drone—contain a trace signature that can lead governments back to the person who recorded it. In other words, human rights activism can shed light on abuses, but it can also yield information about the person publicizing the abuses and possibly put them at risk. As the human rights organization Witness has noted:
It is clear that new technologies, particularly the mobile phone, have made it simpler for human rights defenders and others to record and report violations, but harder for them to do so securely. The ease of copying, tagging and circulating images over a variety of platforms adds a layer of risk beyond an individual user’s control. All content and communications, including visual media, leave personal digital traces that third parties can harvest, link and exploit. Hostile governments, in particular, can use photo and video data—particularly linked with social networking data—to identify, track and target activists within their own countries, facilitated by the growth of automatic face detection and recognition software.90
Human rights activist organizations are now faced with an additional set of challenges if they are to include drones in their arsenal to expose human rights abuses. They must develop protocols to “scrub” video footage of personal data and secure means of holding and transmitting video evidence collected from drones in order to protect the activists themselves.91 They will also have to find some means of authenticating the video and ensuring that it shows what its authors say that it does. This feat is especially difficult in chaotic wartime environments, but particularly important if drones are to be a trusted mechanism for collecting usable evidence of human rights abuses and atrocities.
From targeted killings to surveillance, drones enable all sorts of actors—good, bad, and indifferent—to pursue goals and take risks that they may have otherwise found too costly, difficult, or politically controversial. But drones are rarely neutral in terms of power. In the hands of human rights activists, they are a tool which can be deployed against the secrecy of the state to reveal evidence of abuse, government malfeasance, or negligence. In the hands of the government, they can be deployed against those same activists under the guise of law enforcement surveillance. In this case, drones can be folded into the coercive power of the state, amplifying its capacity to watch its citizens, regulate their behavior, and punish those responsible for challenging them.
In the United States, there is some evidence that drones are extending the surveillance capabilities of the government, though their diffusion has been uneven across the federal government. Aside from the expansion of the use of drones by CBP and DEA, other federal agencies have begun to experiment with camera-equipped drones. For example, the Bureau of Alcohol, Tobacco and Firearms (ATF) spent $600,000 on six drones in 2015, but found that they were largely inoperable for most major tasks.92 The FBI invested in thirty-four drones at a cost of $3 million, but also found that some were ill-suited to their needs, and struggled to train and retain qualified pilots.93The chief problem with much of the expansion of drones across the federal government is that it has been conducted without proper oversight or any transparency, as many drone purchases and use were undertaken with no public notice.94 This is partly because many members in Congress encourage drone sales—sometimes to drive business back to their districts—and allow federal agencies to spend on drones with little direct oversight or accountability.95 This loosening of the rules for the sale and use of drones by federal agencies is also supported by pro-drone caucuses, such as the Congressional Caucus on Unmanned Systems (CCUS), and industry groups, such as the Association of Unmanned Vehicle Systems International (AUVSI). Industry groups in particular have used campaign contributions to encourage members of Congress to authorize spending on drones at the federal level. Only the fact that federal spending on drones is so decentralized and often wasteful has saved the US government from developing an even stronger capability to surveil its own population.
The military has also conducted drone overflights over US territory, though the Pentagon insisted that the drone flights were “rare and lawful.”96 There were fewer than twenty instances of the US military conducting drone missions in US airspace between 2006 and 2015.97 In 2018, the US military flew eleven missions in US airspace, mostly for training and emergency response:98 Reaper and Predator drones were used to monitor wildfires and flooding and also in search and rescue.99 The US military is also exploring the use of drones for dropping parcels of food for hurricane and other disaster relief.100 The Pentagon insists that there is a high bar for approving these missions, based largely on necessity and operational need, and that the Secretary of Defense or someone authorized by him or her must be responsible for approving any proposed mission. At present, the Pentagon does not permit armed drones for anything other than training purposes and does not permit surveillance of US citizens unless authorized by law and approved by the Secretary of Defense.101 But the law and guidelines governing drone use remained somewhat vague and the Pentagon admitted that there were no regulations for how to respond if there was a request from a state or local official (for example, a governor) to “borrow” US military drones in a crisis.
This is a problem because on balance there has been more interest in surveillance at the state and local level. A number of police organizations have been publicly weighing policies to use drones to monitor crowds in protests and at major sporting and entertainment events like the Super Bowl or the Oscars. If this became standard practice, it would effectively make surveillance without a warrant a routine procedure at events with big crowds. But no effort to make this a standard practice has yet succeeded. In 2013, the Los Angeles Police Department was reported to be examining the use of drones to monitor the crowds at celebrity-filled events like the Oscars.102 By 2018, however, they had yet to use drones for this purpose due to sustained opposition by civil liberties groups.103 In May 2018, a bill in the Illinois state legislature would have allowed police to use drones to take photos and videos of protests, which in turn could be used to develop lists of those participating.104 With intense opposition from the ACLU and other privacy advocates, it was ultimately defeated.
One reason why drone surveillance has been so controversial in the United States is that its political implications may be very different. Until recently, surveillance largely occurred with the knowledge or even complicity of those surveilled. The most famous example of extreme surveillance, the panopticon of philosopher Jeremy Bentham, presumed that all involved knew they were being watched and in fact changed their behavior accordingly.105 Even weaker and more limited forms of surveillance in prisons, schools, and factories tended to assume that those being watched knew they were being watched; this fact, and the implied power behind it, had a disciplinary effect on people’s behavior within those spaces.106 The same dynamic applied in public spaces. For decades, individuals in airports and train stations have known that their actions would be captured on CCTV cameras and adjusted their public behavior accordingly.
At the same time, everyone also knew that there were spaces—perhaps at home, perhaps in marginal or neglected parts of each town or city—where no cameras would be present. It was in these unmonitored spaces, often in cafes, theaters, and schools, where political activism and dissent typically flourished. Even in democracies, dissent did not generally flourish in well-monitored spaces where the oppressive glare of surveillance could reveal who was mobilizing and why. Anonymity was an essential precondition to challenging the government. This is why anti-government protesters tended to adopt masks such as the famous Guy Fawkes mask in public protests, even in democracies. In autocracies like the Soviet Union, the preservation of these scarce private spaces—sometimes in university classrooms or cafés—was particularly crucial for providing the anonymity needed to challenge the regime. But today the combined effects of multiple new technologies—from the monitoring of internet traffic and metadata to facial recognition technology and now drones hovering over public spaces—is eliminating anonymity and crippling effective dissent. This explains in part why protesters in the United States so often cite George Orwell’s “Big Brother” when law enforcement propose expanding drone surveillance over protests and other political events. There is a sense that freedom is possible only when some kind of anonymity from surveillance, however limited, is also possible and that many types of modern technology are effectively erasing the possibility of anonymity.
The spread of drone surveillance is dangerous from this vantage point for two reasons. First, drone surveillance may strengthen the hand of government or law enforcement officials against protesters, tilting the balance even more toward those who already have power. Second, drone surveillance can eliminate the anonymity needed to effectively challenge the state and reduce or eliminate those areas in the public and even private sphere when an individual could safely assume they were unmonitored. With drones that can fly and record nearly everywhere with little trace or notice, the area where anyone is truly free from being watched is diminishing by the day. Although the US Department of Justice insists that it will not monitor political protests, many state and local law enforcement officials have not made such a pledge and can use drones, along with other new technologies, to do so.
At present, the degree to which surveillance can identify political protesters and dissenters is partially limited by the technology itself. Despite vast improvements in their cameras, drones cannot see everything on the ground and have a limited perspective on what they do see. But facial recognition technology is now growing by leaps and bounds and will eventually be linked to the cameras with which drones are equipped. Pioneered by the military in places like Afghanistan and Iraq, facial recognition technology is now being deployed by a number of law enforcement organizations to identify criminals and build a database of faces that can be located in a crowd. One project, the Biometric Optical Surveillance System (BOSS), is designed to use algorithms to find faces in a crowd and match them against databases of known criminals.107 But although BOSS was able to make identifications quickly, it struggled with accuracy and issued too many false results, especially with photos taken at an angle. Other programs have been tried by local police departments on an experimental basis, but the overall success rate remains low. The FBI is currently developing a Next- Generation Identification program that will collect fingerprints, iris scans, and facial recognition photographs and allow cross-checking against driver’s licenses and other photos.108This will eventually be handed over to all 18,000 or so law enforcement organizations in the United States. As of 2016, an independent analysis of facial recognition programs found that photographs of 117 million adults in the United States are stored in facial recognition databases. Even though facial recognition software has a high rate of error, particularly against minority groups such as African Americans, relatively few legal and political barriers appear to be standing against its use.1109Safeguards against data misuse are also weak, especially inside the government. In 2019, it was revealed that Immigration and Customs Enforcement (ICE) used facial recognition technology on millions of photos kept in state driver’s licenses databases.110 Over time, facial recognition software will be connected to feeds coming from drones and will make the possibility of identifying the faces of protesters in a crowd a reality.
While protesters in the United States and Europe are reacting to the possibility of Orwellian big brother surveillance of this kind, in other countries it is already occurring. In 2012, Russia began to use drones to monitor pro-democracy protesters.111 Today, the Russia Guard (which is similar to the national guard) uses drone technology stockpiled by the Interior Ministry to watch political protests.112 In China, drones are used to monitor border regions, such as Xinjiang, while the central government is using facial recognition technology to scan people going in to tourist sites and compare them to a centralized criminal database, ironically dubbed Skynet, as in the Terminator films.113 Especially in regions where the persecuted Muslim minority Uighurs live, the Chinese police use a range of biometric data and other means—for example, GPS trackers in vehicles and government spyware loaded into cell phones—to track individuals who may be considered suspect. Citizens are even given a numeric score, which indicates whether they are considered safe or suspect and determines the rights and opportunities they are provided by the government.114 Although drones are only one part of the vast surveillance apparatus that has been constructed by the Chinese state, they have played an important role in diminishing anonymity and limiting the possibilities for dissent in some of the most restive regions of the country. In 2017, only 1,000 drones were deployed for tasks such as tracking police suspects, monitoring traffic, and locating opium farms, but more will come in the future as the intense surveillance of Uighur population increases.115
Another potential danger of using drones to augment the surveillance and coercive capacity of the state is that they may make monitoring, and even repressing, suspect populations easier. This can be seen in the way modern technology has led to the building of internment camps in Xinjiang, but it can also be seen elsewhere. Pakistan has used drones for targeted killings in the FATA. Israel has been using drones for surveillance of Palestinian populations in the West Bank and Gaza for some time but has recently shifted to policing these areas from the air. For example, in March 2018, the Israeli Defense Force (IDF) used drones to deploy tear gas and disperse protesters in Gaza116 and again in May 2018 when mass protests broke out over the decision by the United States to open a new embassy in Jerusalem.117 Against these drones dispersing tear gas, Palestinians flew kites, some of which were equipped with crude explosives. It was the first clear example of what might be described as imperial policing by drones, where a superior government force uses drones to keep a restive population away from their main population centers. As the attacks continued, Israel responded with drones that could knock down the flaming kites and even drop tear gas along the border.118 For the Israeli government, drones are helping to make a morally intolerable situation more politically acceptable by reducing casualties at home and keeping the conflict out of sight of most of the population.
This practice is spreading. As of 2019, both Pakistan and Nigeria have conducted drone strikes on secessionist regions, suggesting other governments will look to drones to manage problems that would otherwise demand a political solution. The risks of such an approach are growing as drone technology becomes cheaper and more widely available. Although Israel used its own drones for its missions, a number of private companies are now offering crowd control drones at cheap rates on the international market. One drone, the Skunk riot control copter, is equipped with tear gas grenades, paint and pepper balls, and a set of loudspeakers to communicate with the crowd.119 It is being sold by Desert Wolf, a South African defense company, and has found an eager market among some police and private companies. Other similar, models, equipped with cameras and riot control agents like tear gas and smoke bombs, are now being developed or sold by Chinese, Israeli, and Austrian companies.
It is commonly assumed that a world of drone surveillance will resemble some kind of science fiction nightmare, with governments using drones to convict people of crimes they have not yet committed, as in the film Minority Report. Instead, the dangers of the spread of surveillance drones are perhaps less obvious but just as worrying. The use of surveillance drones is growing, even if unevenly, across a range of law enforcement and government actors whose appetites for knowing what their populations are doing is near insatiable. This is happening in an unchecked, non-transparent way with little government direction or oversight. From the adoption of the PSS by Baltimore to the spread of small surveillance drones to dozens of small state and local police forces, surveillance is increasingly becoming normalized across the United States and Europe. A range of actors—state and local law enforcement, but also private companies—are eager to deploy drones and see what they can see. Ordinary legal protections of privacy have proven ineffective against the spread of the surveillance state, both through drones and other technology like CCTV. Given their flexibility and ability to reach less covered spaces, there is a real danger that the diffusion of surveillance drones may reduce the sphere of privacy needed for effective political mobilization and dissent. If drones are ultimately put to watching and identifying political protesters using facial recognition technology, as currently proposed, they will have significantly amplified the coercive power of the state against citizens and also corroded the democratic character of those governments.
Outside the United States, drones are being added to the arsenals of authoritarian governments like Russia and China. Here the danger is greater: drones will eventually become an essential component of the growth of their modern surveillance state, recording protesters in public areas and keeping watch on previously private areas as well. In this way, drones will tilt the balance of power toward the government and make protest and dissent by citizens riskier than it is today. One consequence of the expansion of drone technology may be to make the strong even stronger, thus shifting the balance of power decisively toward those who wield the coercive instruments of power and against those who dare to challenge them. Further, if drones become a tool of imperial policing by states like Israel, there are additional dangers. As the technology develops, we can expect to see more drones equipped with tear gas, smoke bombs, and even small explosives that will be able to disperse crowds and punish those who challenge the existing political order. If drones become a tool of political repression and allow governments to police their domestic enemies with ruthless efficiency, governments with secessionist movements might be less willing to negotiate and grant concessions. The result might be that such conflicts are contained, but not resolved, while citizens in developed states grow increasingly indifferent to the suffering of those making secessionist or even national liberation claims, against them, even when these claims are just. If the gaze of surveillance drones increases the pool of information gathered about suspect populations but not the level of understanding or sympathy for them, it will be hard to see drones as a net positive for human society, especially when concentrated in the hands of the strong and privileged.