Police Surveillance and the Perils of Techno-Solutionism

Prithika Balakrishnan September, 2026

Introduction

Generalized, routine surveillance has become an unavoidable hallmark of daily life in the United States. Americans are caught on public camera over 200 times a week on average.¹ The scope of this surveillance is especially notable in the context of policing.² As of 2024, 1,531 police departments operate video-capable drones, 1,722 employ license-plate readers, and 898 utilize facial recognition technology.³ These sources of surveillance data are consolidated by hundreds of police departments in Real Time Crime Centers.⁴ Furthermore, Americans are increasingly participating in their own surveillance by using smart doorbells and motion-sensing security cameras.⁵ A growing number of local police departments incorporate these private cameras into their surveillance nets, expanding their camera capacity by the thousands.⁶ “The scope and variety of the types of surveillance that are possible today are unprecedented in human history.”⁷

These statistics, however, do not answer fundamental questions about the purpose of law enforcement surveillance. Critically, what specific problem is police surveillance intended to fix? Is it being used as a tool for deterrence? Evidence gathering? Crime prediction? There is nothing inevitable about the way that American law enforcement uses surveillance technology. Not only have police departments made choices as to which surveillance tools to employ and to what predetermined ends, these choices have had consequences beyond their stated objectives. The impacts of technological tools are not easily cabined. “[T]echnology’s interaction with the social ecology is such that technical developments frequently have environmental, social, and human consequences that go far beyond the immediate purposes of the technical devices and practices themselves . . . .”⁸ Scholars have elaborated upon this narrative, pointing to the ways in which facially neutral law enforcement technologies, built as they are upon systems of racial and other hierarchies, result in racially biased and otherwise unjust outcomes.⁹

The story, however, does not stop there. Technology does not simply meet existing goals or have broad consequences beyond its initial goals. It also alters the end goals themselves. As sociologist Gary T. Marx states, “New means, beyond meeting old needs and appearing in response to new needs, may play an independent role. In considering questions of invention, whether social or material, new means may generate new needs and in a seldom-recognized process may even determine ends . . . .”¹⁰ While a tool may be used to respond to a specific problem, it also changes the nature of the problem itself and alters the array of potential solutions. Further, technological tools are not “merely passive instruments” and “may actively influence their users, changing the way they perceive the world, the way they act in the world and the way they interact with each other.”¹¹ Far from being a neutral tool, technology has been re-imagined by philosophers as “[a] moral agent[],” with its own orientation and momentum.¹²

This Article brings these historically siloed scholarly traditions into conversation. It asks not only how surveillance has met existing policing needs, but how it has changed law enforcement goals and the nature of law enforcement itself. While legal scholars have discussed the Fourth Amendment implications of law enforcement surveillance,¹³ and the racialized implications of police technology,¹⁴ this work extends this analysis by focusing on how surveillance technology has changed policing goals, with racialized implications that go beyond the reach of the Fourth Amendment.

This paper frames law enforcement’s embrace of surveillance technology as a form of techno-solutionism, defined as the view that all complex social situations can be recast “as neat problems with definite, computable solutions or as transparent and self-evident processes that can be easily optimized . . . .”¹⁵ It is the belief that technology can offer quick-fixes and magic bullets to solve a wide range of problems in education, urban planning, public health, climate change, poverty, policing, and more.¹⁶ It is a mode of solving problems that reaches “for the answer before the questions have been fully asked.”¹⁷

Part I starts with a recent example of this larger techno-solutionist narrative in law enforcement. As police departments around the country have faced recruitment struggles in the 21st century (whether real or perceived),¹⁸ surveillance technology has served as a presumptively apolitical solution to the problem of fewer officers. This techno-solution to police staffing challenges uses surveillance technology as a “force multiplier,” promising increased efficiency in meeting existing policing goals with fewer officers.¹⁹

The expansion of surveillance technology into local law enforcement to address staffing challenges is a recent example of a larger technosolutionist narrative in American policing. This trajectory originated during the backlash to the Civil Rights movement in the late 1960s and has continued through the War on Terror to the current moment. Part II charts this larger history by using closed-circuit TV (CCTV) as a case study, showing how the techno-solutionist mindset interacted with three overarching historical and technological developments and ultimately changed policing goals. First, federal funding, born from the racialized politics of the War on Crime and nourished during the War on Terror, encouraged data coordination and crime prediction. Second, corporate partnerships obscured concerns about surveillance’s efficacy and racial bias through their focus on profit and market growth. Third, the explosion of participatory surveillance resulted in the disparate spread of technology along racial and socio-economic lines.

Part III focuses on how expanded police surveillance, unmoored from historical and socio-economic context by techno-solutionism, has strengthened existing patterns of biased policing through racially patterned and pervasive surveillance. The expansion of police surveillance in the last fifty years, in general, and its techno-solutionist trajectory, in particular, lead to troubling implications about how constant surveillance will impact racial bias in the criminal legal system. The vast amounts of information harvested from ubiquitous surveillance have culminated in the creation of “suspect data doubles,” dossiers that reduce each of us to the uninterrogated correlations mined from boundless data. The era of Big Data policing, now upon us, is positioned to alter law enforcement goals and amplify racial disparities under the cloaked passivity of technology.

I. The Problem and Its Techno-Solution

Part I explores a recent example of the broader techno-solutionist narrative in law enforcement. As police departments across the country have struggled with real and perceived recruitment issues in the 21st century, surveillance has been presented as a neutral and unbiased technological workaround. This techno-solution to staffing shortages uses surveillance as a “force multiplier,” promising greater efficiency in achieving policing objectives with fewer officers, while sidestepping long-standing sociopolitical critiques.

A. Police Staffing Challenges

Why have police departments turned to surveillance technologies in the 21st century? A clue may be found in one of policing’s top concerns – staffing. Since the early 2000s, law enforcement has faced challenges recruiting and retaining officers.²⁰ In 2009, the RAND Corporation’s Center on Quality Policing reported that, “Maintaining the police workforce level is continually one of the greatest challenges facing law-enforcement agencies.”²¹ By 2018, “a clear and worrying trend” had emerged in police recruitment and retention.²² Approximately one quarter of police departments reported staffing difficulties in 2018, which increased to 78% by 2022.²³ In 2024, the Police Executive Research Forum (PERF) reported that, in California, law enforcement staffing was at its lowest since 1991, despite a population growth of 10 million during the same time period.²⁴

Police staffing challenges have unfolded as the result of a web of interconnected recruitment, retention, and retirement issues, and they grew to a crescendo in the early 2020s. Difficulties with recruitment have emerged as fewer people apply to become police officers. A September 2019 survey by the International Association of Chiefs of Police found that 78% of agencies reported “having difficulty in recruiting qualified candidates,” with 75% reporting that “recruiting is more difficult today than it was five years ago.”²⁵ For example, the Nashville Police Department reported a 60% decrease from 2010 to 2019 in the number of online applications it received.²⁶ In 2023, the New York State Police had only 12,000 applicants take the required agency exam, down from over 30,000 a few years prior.²⁷ San Francisco, where police officer application numbers had already been dropping since 2014, saw its numbers fall to a quarter of the previous levels in 2021.²⁸

While staffing challenges are not new, they have been heightened by the “twofold morale hit of 2020 – the coronavirus pandemic and criticism of police that boiled over with the murder of George Floyd by a police officer.”²⁹ This effect was found to be highest for resignations and retirements, which have increased 19% and 47% respectively, as compared to 2019.³⁰ Studies report a widespread “turnover crisis,” as 78% of policing agencies experienced increased resignations and retirements from 2020 to 2022.³¹ Total resignations of sworn police officers increased by 2.5% from 2019 to 2020 and then by almost 40% in 2021 and 14% in 2022.³² While the pace of resignations slowed in 2023, the numbers are still 28.6% higher than 2019 levels.³³ Officer retirements also increased from 2019 to 2020, and again in 2022.³⁴ A study of fourteen large police agencies found that 80% had experienced increased resignations and retirements in the two years following the George Floyd protests.³⁵ The combination of increased retirements, earlier resignations,³⁶ and smaller hiring pools has resulted in a widespread decrease in the number of police officers. In California, staffing fell by 3.2% from 2020-2022, with a loss of 3,563 sworn police officers statewide.³⁷

The causes and extent of these post-2020 staffing challenges – and the actual impact of public criticism of law enforcement based on well-publicized police murders of Black and Brown constituents – have been debated by scholars. Several empirical studies point to substantial law enforcement turnover in the wake of George Floyd’s death.³⁸ Other studies find that the effect was minimal, short-lived, or due to other factors.³⁹ Regardless of whether these post-2020 staffing challenges were real or exaggerated, media coverage and law enforcement statements created a perception of increased police retention and recruitment difficulties.⁴⁰

Alongside staffing challenges are changes in public opinion of the police. Gallup polling began measuring confidence in police in 1993, with a high mark in 2004, when 64% of respondents expressed “a great deal/quite a lot” of confidence in police. This number fell to 48% in 2020 and again to a record low of 43% in 2023, before rebounding to 51% in 2024.⁴¹ According to ABCNews/Washington Post polling, in 2014, 54% of respondents expressed confidence in the adequacy of police training to avoid excessive force. By 2023, that number had fallen to 39%.⁴² Similarly, from 2014 to 2023, the percentage of respondents who agreed that police treat Black and white people equally fell from 52% to 41%.⁴³ In a different poll, the percentage of respondents who viewed police violence against the public as “extremely/very serious” grew from 32% in 2015 to 48% in 2020.⁴⁴

The fall in public confidence and the difficulties in maintaining police staffing have been portrayed as an existential dilemma. The U.S. Senate Judiciary Committee held a hearing in 2015 entitled, “The War on Police.”⁴⁵ In August 2023, PERF identified recruitment and retention as the profession’s greatest challenge.⁴⁶

While the causes and extent of the staffing crisis are varied and debatable, the broader trends point to fewer Americans vying for law enforcement jobs and a decline in the public perception of police. As Sam Blonder, CEO of a police recruiting agency, stated, “Ask 150 high school kids who wants to be a police officer – you won’t get one that will raise their hand. It’s not for me to say why that’s happened, but sometimes an industry needs a shakeup like that.”⁴⁷ Although responses to such a “shakeup” could theoretically include strategies such as addressing police brutality, reducing implicit bias among law enforcement, diversifying cadet recruitment, deploying non-police first responders to certain incidents, or altering the relationships between law enforcement and communities of color,⁴⁸ many police departments have instead turned to surveillance technologies to solve staffing challenges. This move towards a technological answer, as opposed to “problem-driven design,”⁴⁹ is an example of techno-solutionism at work.

B. “Force Multipliers” and Techno-Solutionism

To many surveillance industry leaders and police chiefs, the answer to fighting crime effectively with fewer people is technology. “The good news,” according to industry magazine Security Today, is “that many are discovering how advances in video surveillance technology can be an affordable force multiplier that delivers effective results and safer cities.”⁵⁰ According to Motorola, “wireless video’s ability to monitor multiple locations roundthe-clock serves as a significant force multiplier, helping officers and other department employees become substantially more productive.”⁵¹

The term “force multiplier,” originating from military science, refers to “a tangible or intangible variable that increases the combat value and overall capability of a military force.”⁵² It has become a buzzword in law enforcement circles, where it describes the increased efficiency made possible by technologically enhanced policing.⁵³ “By leveraging technology as a force multiplier, you enhance the efficiency and precision of your officers.”⁵⁴ In Camden, New Jersey, Police Chief Gabriel Rodriguez welcomed an infusion of $8 million for the Eye in the Sky surveillance system by stating, “I wish we had a thousand officers on every corner, but we don’t,” and calling the technology a “force multiplier.”⁵⁵ In touting the capabilities of integrated surveillance technologies, Marshall Freeman, Deputy Chief Administrative Officer for the Atlanta Police Department, stated, “[Real-Time Crime Centers’] simplicity and effectiveness make them a force multiplier for enhancing efficiency and safety.”⁵⁶ Technologies such as automated license plate readers and unmanned aerial vehicles “can often be implemented at the cost of less than one full-time benefited police officer.”⁵⁷

Through the lens of force multipliers, local police departments’ turn to increased surveillance can be seen as more of a response to staffing shortfalls than a strategic shift to a technology with proven results in reducing crime. Essentially, crime isn’t actually the problem that technology is being recruited to solve. Rather, it is being used to solve issues internal of policing itself. How can police departments do the work expected of law enforcement with fewer employees, given that the ability to retain existing officers and recruit new ones is waning? In the words of Memphis Police Chief Cerelyn Davis, “We are looking forward to more advanced technology, more cameras, more equipment for our officers to work smarter, not harder.”⁵⁸ Smarter, here, is a synonym for efficiency.

This turn to technology is a brand of “techno-solutionism,” which is the view that all complex social situations can be recast “as neat problems with definite, computable solutions or as transparent and self-evident processes that can be easily optimized . . . .”⁵⁹ It is an ideology that refers to “the human tendency to place emphasis upon a specific solution to a problem rather than on the problem itself.”⁶⁰ While techno-solutionism is not unique to the field of policing – education⁶¹ and healthcare⁶² also grapple with the allure of new technology as a panacea for persistent dilemmas – it can be seen prominently in many aspects of policing. One example can be found in the ideas of Axon Enterprise CEO Rick Smith.⁶³ Smith’s company (formerly Taser International) sells new technologies, like Tasers, as non-lethal alternatives to firearms for police. Smith envisions a world in which killing is obsolete: “Killing is a technology problem. And we can use technology to end killing as we know it.”⁶⁴ According to Smith, the obstacles to a safer world are not persistent inequalities, but rather our skepticism of new technologies like crowd control, surveillance, and artificial intelligence.⁶⁵

Most recently, techno-solutionism has been applied to police staffing. Here, the identified problem is a dearth of police officers, and a typical solution would be to recruit more police officers. When that typical solution is no longer possible, rather than approaching the problem of crime differently – with mental health outreach, educational opportunities, or a larger socio-economic reckoning, for example – or assessing the root causes of why a career in law enforcement has become less attractive, techno-solutionism promotes technology as the solution instead. Under this approach, the answer to the shortage of police officers is to increase technological surveillance.

This techno-solution is aided by law enforcement’s internal policymaking and public perception. First, the decision-making process at the local police level is remarkably unconstrained. “In most cities, decisions to acquire and use spying technologies are made by police departments, in secret, without any knowledge or input from the public or their elected officials.”⁶⁶ Most jurisdictions do not have clear surveillance policies, leaving it to law enforcement to make decisions based upon their own rubrics.⁶⁷

Further, within local police departments, “technology acquisition seem[s] rather disparate and [is] not necessarily obtained for part of a greater purpose.”⁶⁸ A 2016 Department of Justice funded research report found that, “at a national level, agencies are not making decisions to acquire technology based on dominant policing philosophies or the activities they prioritize. Instead, agencies appear to adopt technology ad hoc in response to a constellation of factors that includes executive staff decisions, perceived needs, community demands, and available funding.”⁶⁹ The availability of federal funding, private citizen interest, vendor sales-pitches, and the variability of the political climate leads to “ad hoc” decisions with little foundation in real strategy.

Finally, perceptions about crime, not realities, have had an outsized impact on policing strategies and policies. Although crime statistics are largely trending downwards,⁷⁰ there is nevertheless increased public concern over crime. Gallup polling numbers from October 2023 found that 77% of respondents believed, incorrectly, that crime rates had increased in the last year.⁷¹ The percentage of respondents who believed that “the crime problem in the U.S. [was] either extremely or very serious” had jumped from 54% in 2021 to 63% in 2023, the highest ever recorded by Gallup.⁷² Further, a “growing share of Americans [said] reducing crime should be a top priority for the president and Congress to address [in 2024],” with that share increasing from 47% at the beginning of 2021 to 58% in early 2024.⁷³ These factors have facilitated the spread of surveillance as a “technical” solution to the socio-political problems of policing, which include racial bias, public distrust, and police brutality.

The issue is not that technology might indeed offer tools to address entrenched social problems,⁷⁴ but that techno-solutionism professes an often unquestioned “numeric imagination:” a “predisposition to seek out quantitative and linear casual explanations that have little respect for the complexity of the actual human world.”⁷⁵ Techno-solutionism can obscure the need for structural change or analyses of complex socio-political dynamics. In the process of identifying problems, they are “rendered nonpolitical.”⁷⁶ The underlying issues with policing, however, are indeed socio-political ones, where context and histories of power matter.

C. Police Surveillance and Public Safety

Techno-solutionism frames human concerns as technical problems, often ignoring broader structural or socio-political complexities. However, a critique of techno-solutionism need not deny that technology can offer tools to address underlying problems. One way to analyze the efficacy of police surveillance is to measure its effects on crime. This approach assumes that crime is at the heart of why police have turned to surveillance, which is the rationale given by law enforcement and policymakers pushing for greater surveillance with fewer restrictions. For example, Los Angeles Deputy Chief John McMahon stressed that, “The potential of video technology to impact law enforcement significantly, especially in real-time, cannot be understated . . . By enabling immediate access to video and other crucial data for officers and detectives on the scene, we can significantly enhance our crime-fighting capabilities.”⁷⁷ Taking the argument that technology is an effective tool against crime as our starting point, how well does police surveillance serve the goal of reducing crime? The answer is “it depends,”⁷⁸ as we will discuss in this section.

Answering a complex, multi-factor question, such as the effect of law enforcement video surveillance on crime rates, is a precarious task. An honest but unsatisfying response is that more data is needed. Looking solely at macro-level statistics, a comparison of city crime indices, and the number of cameras per 1,000 people, there is little correlation between more cameras and lower crime rates.⁷⁹ Atlanta has six times the number of police cameras per capita, as compared to Denver and Detroit.⁸⁰ However, Denver’s crime rate is substantially lower than Atlanta’s, while Detroit’s crime rate is substantially higher.⁸¹ These types of comparisons tell us little about the efficacy of police surveillance, as the factors that make up the crime index of any given jurisdiction are impenetrable without a more detailed analysis.

Another, albeit imperfect, approach might be to evaluate a single jurisdiction, comparing pre-surveillance and post-surveillance crime outcomes. Again, this approach is hampered by the fact that few law enforcement entities track this type of information in any consolidated fashion. A 2016 Department of Justice survey of police departments reported that “agencies largely do not identify or track appropriate success metrics that could be used to determine the effectiveness of technology.”⁸²

One study conducted by the University of California in 2008 offers an early snapshot of the effectiveness of San Francisco’s initial surveillance project, the Community Safety Cameras (CSCs).⁸³ At the time, the CSC program, which had originated in 2006, had seventy-one fixed view cameras, mostly positioned in the Tenderloin, Mission, and Bayview Districts.⁸⁴ The study noted two primary goals of the program: first, to deter violent crime, and second, “to provide SFPD investigators with a forensic tool to investigate crime that occur[s] within view of the cameras and to assist in the prosecution and defense of charged crimes.”⁸⁵

With regards to the first goal, the study stated:

We find no evidence of an impact of the CSCs on violent crime. Violent incidents do not decline in the areas near the cameras relative to areas further away, we observe no decline in violent crimes occurring in public places, and we observe no statistically significant differences in the relationships between the before-after change in crime and the distance from the camera locations.⁸⁶

The study also did not find evidence of any effect on drug offenses, prostitution, or vandalism.⁸⁷ It did measure, however, “significant and substantial declines in property crime rates within view of the CSCs.”⁸⁸

With regards to the second goal, the study did find that CSC footage was helpful in criminal investigations, specifically with “establishing a sequence of events for a crime or placing witnesses at a scene.”⁸⁹ On average, SFPD had requested CSC footage three times a month since the program’s inception, and it had assisted in charging suspects in six cases overall.⁹⁰ Nevertheless, the report concluded, “given the lack of deterrent effect on violent crime and its limited usefulness with respect to investigations, it is clear that one of the questions the City must consider is whether or not to continue the CSC program at all.”⁹¹

In the years since the 2008 University of California report, San Francisco made the decision to not only continue, but to expand law enforcement surveillance. In 2022, San Francisco launched a pilot project allowing for real time police video surveillance.⁹² The results of the pilot project over a nine month period show that live monitoring led to sixty-five total arrests, with the majority (fifty-three) being for narcotics-related incidents.⁹³ Of these arrests, twenty-nine led to charges being filed.⁹⁴ Of the 250 hours of live video watched by SFPD officers, forty-two hours and “by far the longest monitoring session” was of a music festival in a public park.⁹⁵ The surveillance netted only one arrest.⁹⁶

These underwhelming measures of surveillance efficacy are not unusual. The Memphis Police Department’s extensive surveillance system, Skycop, mirrors the outcomes in San Francisco. In the aggregate, violent-crime rates in Memphis have increased by 57% since 2010, the year Skycop was implemented, outpacing Tennessee and national averages.⁹⁷ Would the violent crime rate in Memphis – which is six times the national average and climbed at a higher rate in the 2010s than anywhere else in the country – have increased even more absent Skycop cameras?⁹⁸ Given the myriad factors that converge to create crime rates, answering this hypothetical conclusively is difficult. However, Memphis’s experience with surveillance as a crime fighting tool is matched by other jurisdictions, which have also measured few or inconclusive effects from increased law enforcement surveillance.⁹⁹ In the words of former Memphis Police Chief Don Crowe, “We tell everybody this anecdote: every bank has cameras in it, but people still rob banks.”¹⁰⁰

What can be more easily measured is the rate at which Skycop surveillance was used by law enforcement to respond to crime incidents. Based upon a public records request by the Daily Memphian newspaper, from January 1 to October 31, 2021, Skycop footage was mentioned in only one of the 228 homicides in the city.¹⁰¹ Skycop surveillance was also mentioned in less than 5% of the aggravated assaults, carjackings, and robberies during that same period.¹⁰² Overall, it was relevant to less than 3% of all crime incidents in the city.¹⁰³

Similar results can be found in Chicago. A joint study by the Chicago Tribune and the Illinois Answers Project based on over 100 open records requests found that, “videos from Chicago police’s core inventory of roughly 4,400 surveillance cameras appeared to help solve, at best, 3.5% of 2023’s homicides.”¹⁰⁴ Despite costing almost $300 million over the course of two decades, only about one in ten cameras documented a significant drop in crime.¹⁰⁵

Interestingly, given their role in promoting visual surveillance, corporate security companies offer few statistics extolling the efficacy of their products. The websites of companies like Fusus or Motorola are replete with anecdotal accounts, testimonials, and case studies, but sparse on crime data.¹⁰⁶ One of the few companies to offer a detailed success story is Genetec, which quotes the Mayor of Detroit in 2017, crediting Project Green Light (PGLD) with reducing violent crime at participating gas stations.¹⁰⁷ However, as noted in a 2020 Michigan State University (MSU) report, PGLD “was implemented during a period of multiple crime and violence reduction strategies,” which were collectively effective in reducing crime.¹⁰⁸ “[I]t becomes very difficult in distinguishing the impact of any one of these strategies relative to the other strategies.”¹⁰⁹ Overall, the report found “limited impact of PGLD participation on trends in crime.”¹¹⁰

Like the MSU report, a growing body of literature has focused on the efficacy of technology as a crime deterrence tool. A 2009 study conducted a systematic review of academic studies on CCTV use by law-enforcement. It found a 16% decrease in crime post-surveillance installation, a figure which was driven by “the effectiveness of CCTV schemes in car parks, which caused a 51% decrease in crime. Schemes in most other public settings had small and nonsignificant effects on crime . . . .”¹¹¹ A 2019 meta-analysis including seventy-six “CCTV evaluations published from 2007 through 2017” found that, overall, “CCTV was associated with significant reductions in both vehicle crime and property crime in general, with no significant effects observed for violent crime. Public safety agencies combatting violent crime problems may need to consider whether resources would be better allocated toward other crime prevention measures.”¹¹² Overall, the common thread in evaluations of police camera surveillance is that, “[t]hough findings are mixed, they indicate cameras can reduce crime, particularly property crimes and vehicle crimes in parking lots”¹¹³ but have “no violent crime reduction effects . . . .”¹¹⁴

Despite these lukewarm surveillance results, local law enforcement and political leaders continue to laud technologically aided police surveillance as a necessary tool for public safety, often raising the specter of violent crime. When Amazon limited the ability of law enforcement to directly contact Ring doorbell customers via the app in early 2024, Fresno Police Chief Paco Balderrama reacted with alarm: “It’s gonna delay. When we have a known killer, a known rapist, a known gang member that just shot somebody, and we’re trying to get them off the streets, it’s critical that we get that information to detectives immediately.”¹¹⁵ In 2022, the San Francisco Mayor’s office argued that increasing police department access to real-time cameras would allow law enforcement “to respond to the challenges presented by organized criminal activity, homicides, gun violence, officer misconduct, among other crimes . . . .”¹¹⁶ When describing the benefits of the Hendersonville, Tennessee network of surveillance cameras, Police Commander Scott Ryan stated, “[W]e want our protectors in the field to have information instantly in order to provide the most effective and the safest response to violent crimes . . . .”¹¹⁷ At times, the push for increased surveillance seems to be an end in and of itself. In the words of Houston Police Chief Troy Finner, “I want thousands [of cameras] . . . Thousands. The more, the better.”¹¹⁸

Despite the lack of evidence showing that police surveillance impacts violent crime rates, law enforcement and politicians continue to extoll its value in combatting dangerous criminals. This enthusiasm has facilitated the spread of surveillance as a “technical” solution to the socio-political problems of policing. The ways that technology has been invoked as a solution to police officer shortages is part of a larger trend towards techno-solutionism.

II. Case Study: The Ascendance of Closed-Circuit Television

Part II desribes police surveillance technologies in the political, economic, and social conditions of their ascendance. Focusing on law enforcement’s use of CCTV as a case study, it charts how technosolutionist thinking undergirds the adoption of technology by police, as well as how technology has changed the goals of policing from deterrence and solving crimes to data aggregation and prediction. Throughout, the specific American experience of race surfaces repeatedly. Police surveillance first emerged from 1960s War on Crime-era racialized politics, benefiting from new federal funding for local law enforcement. It gained momentum through the War on Terror, reorienting local police to data integration, crime prediction, and surveillance. Corporate partnerships emerged, prioritizing profit over concerns about technology’s collateral effects. Finally, the growth of participatory surveillance allowed for trends which emphasized economic and racial disparities. The history of law enforcement CCTV offers a view into the factors which played decisive roles in the development of surveillance in the United States. It also affords a clear vantage point from which to observe the interplay between technosolutionism and the continuation of racially biased policing.

A. Federal Funding

1. War on Crime: Racialized Roots of Federal Funding

American police first employed CCTV surveillance in September 1968, on the main business street of Olean, New York.¹¹⁹ The local Sears store had been repeatedly broken into, and company headquarters had ordered the store manager to purchase a guard dog that he would have to care for, much to his chagrin.¹²⁰ At a community service club meeting, the regional cable television company representative suggested placing police cameras on the main street’s utility poles. The initial system was rudimentary by 21st century standards – eight cameras “pann[ing] left and right on a timer,” with overlapping fields of view that were fed continuously to a bank of monitors at the local police department.¹²¹ The goals were straightforward: deter crime and catch it in action. The system cost $1.4 million, took two years to develop,¹²² and required a police officer to watch the continuous feed in real time, since there was no recording capability.

Olean’s experience as a pioneer in police CCTV was short-lived. The CCTV experiment only lasted a year.¹²³ The New York Times explained that “officials decided there was too little street crime in the upstate city to justify the expense.”¹²⁴ Despite the system’s limitations, it drew national interest from law enforcement: over the next few years, over 160 police chiefs from around the country visited the town of Olean.¹²⁵ This interest, importantly, coincided with two critical technological developments. First, in 1971, Sony introduced the first commercially available VCR, providing momentum for broad commercial and law enforcement expansion of CCTV surveillance.¹²⁶ Users could now record security footage for later review and easily replace VHS tapes for continual surveillance. A human being was no longer required to monitor the cameras in real time. Second, newly developed camera technologies, such as the charge-coupled device, improved image quality and the ability to record in low-light situations.¹²⁷ These technological advances allowed security cameras to become increasingly commonplace in the 1970s.

They also coincided with a seismic shift in federal involvement in local policing. Up until the late 1960s, crime had been seen almost exclusively as a state and local concern. “In the 1950s, for example, there was virtually no federal involvement in the kinds of criminal justice functions that normally and traditionally had been, up until that time, left to state and local governments.”¹²⁸ This changed in 1965, when the Law Enforcement Assistance Act “establish[ed] a direct role for the federal government” in local police operations, court systems, and state prisons.¹²⁹ The Act created the Law Enforcement Assistance Administration (LEAA), “the first program of significant federal aid for state and local law enforcement.”¹³⁰ LEAA was expanded into a formal Department of Justice agency by the Omnibus Crime Control and Safe Streets Act of 1968,¹³¹ as part of the Johnson administration’s attempt to wage “a thorough, intelligent, and effective war against crime.”¹³² LEAA continued to exist until 1980, spending $8 billion in federal funds to support local law enforcement.¹³³ “[A]t its zenith, in 1974, the LEAA made up nearly 45 percent of the DOJ’s spending.”¹³⁴

A study by the United States General Accounting Office in 1977 detailed LEAA funding programs from 1969 to 1970, noting that “Congress enacted the Omnibus Crime Control and Safe Streets Act of 1968 at a time when serious crime was accelerating at an unprecedented rate.”¹³⁵ What the report left out, however, were the other factors in the 1960s that turned local crime into a national concern, paving the way for federal action. The 1960s was a decade that included the assassinations of John F. Kennedy, Martin Luther King, Jr., and Robert F. Kennedy; student protests against the Vietnam War; the nonviolent civil rights movement; and the “long, hot summer” of 1967, when more than 100 urban communities were sieged by violence.¹³⁶ The Omnibus Crime Control and Safe Streets Act was passed the following year, in 1968.

While violent crime did increase by 126% from 1960 to 1970 and by another 64% from 1970 to 1980,¹³⁷ Americans’ fear of crime “increased considerably faster than the crime rate . . . ,” and was not only affected by crime rates, but by the racial composition and proportion of interracial crime in any given city.¹³⁸ A 1982 regression study based on data from twenty-six American cities found that, for whites, fear of crime was increased by property crime rates and the proportion of interracial white victimization.¹³⁹

These fears, when harnessed by politicians, fueled racially coded “tough on crime” rhetoric and policies. “A candidate who declared himself tough on crime was understood to also oppose radical shifts in the social order.”¹⁴⁰ In the 1964 presidential election, Republican candidate Barry Goldwater raised the specter of “crime in the streets” as a national concern.¹⁴¹ In the 1968 election, Richard Nixon successfully raised this flag again, with his law-and-order platform.¹⁴² John Ehrlichman, President Nixon’s domestic policy advisor, admitted in 1994 that the roots of the War on Crime strategy “had two enemies: the antiwar left and black people.”¹⁴³

With federal funding came military technologies and strategies, such as the introduction of Special Weapons and Tactics (SWAT) units into local policing.¹⁴⁴ The War on Crime increased the militarization of local law enforcement in terms of both material equipment and strategy. Federal money facilitated the purchase of “bulletproof vests, helicopters, tanks, rifles, gas masks and other military-grade hardware for police departments.”¹⁴⁵

This funding also facilitated police surveillance. In 1971, LEAA began funding CCTV surveillance projects in cities around the country.¹⁴⁶ By providing grants to early CCTV systems and military technologies, LEAA ushered in a change in domestic policing: an increased reliance on technology. Among its many grants, LEAA also provided direct funding to local police to promote multi-jurisdictional cooperation, including computerized data sharing among disparate government agencies. ¹⁴⁷

The 1990s saw a second wave of CCTV police surveillance, as cities began installing video camera systems for law enforcement purposes.¹⁴⁸ This expansion also coincided with technological developments. In the 1990s, CCTV systems transitioned from analog to digital, allowing for higher quality pictures, zoom capabilities, and easier footage retention.¹⁴⁹ Furthermore, digital “multiplexing” became feasible, allowing a single screen to display multiple camera recordings for more efficient monitoring, and allowing the retention of large amounts of video footage on smaller hard or flash drives.¹⁵⁰ Still, physical access to the surveillance monitor or the storage drive was required.

This, of course, changed with the advent of the internet. CCTV systems incorporated IP cameras, which allowed for surveillance footage to be sent and received over the internet.¹⁵¹ In 1996, Axis Communications launched the first network camera, announcing that, “[f]rom that moment on, people with an internet connection could watch what was going on from anywhere in the world.”¹⁵²

In the thirty years between the emergence of rudimentary CCTV and the broad use of networked cameras at the end of the 20th century, surveillance technology had firmly imprinted itself on domestic law enforcement. Initially envisioned as a tool to serve traditional deterrence and evidence gathering purposes, police surveillance was rapidly evolving, due to technological advances and federal funding. With its origins in War on Crime racialized fears, surveillance technology was positioned to change the goals of policing.

2. War on Terror: A Move Towards Crime Prediction

Like the War on Crime, the War on Terror provided local law enforcement with extensive federal funding for surveillance technologies. It promoted national security strategies and methods, such as networked data gathering and multi-jurisdictional coordination. Moreover, it translated counter-terrorism goals of prevention and prediction into local law enforcement priorities. Federal funding of surveillance has played a part in “the changing role of the police as ‘knowledge workers’ within a society gravitating toward the estimation and prevention of risks.”¹⁵³

After the September 11, 2001 attacks, the FBI underwent a “dramatic new organizational shift,” receiving increased funding to transform itself “from a primarily crime-solving entity to a pro-active domestic intelligence agency.”¹⁵⁴ Numerous urban police departments joined the Department of Justice’s National Joint Terrorism Task Force (JTTF). “[O]ver one hundred JTTF squads [were] directed by the FBI’s field offices and tasked with preventing terrorism and investigating potential terrorists.”¹⁵⁵ In addition to funding, local law enforcement received guidance “as to what type of information they should collect and pass on to the FBI field offices or to the JTTFs.”¹⁵⁶ In San Francisco, for example, local police assigned to the JTTF participated in undercover surveillance of anti-war rallies in the lead up to the second Iraq War.¹⁵⁷ They attended “meetings at mosques and other community organizations under the guise of ‘community outreach,’ while recording information about the attendees’ political and religious affiliations . . . .”¹⁵⁸ “What the War on Drugs [had] done for police militarization, the War on Terror [was] now doing for police intelligence gathering . . . .”¹⁵⁹

The Implementing Recommendations of the 9/11 Commission Act of 2007 funded Fusion Centers throughout the United States.¹⁶⁰ Fusion Centers serve as coordinated hubs for “the receipt, analysis, gathering, and sharing”¹⁶¹ of “critical information housed in different agencies” in order to aid federal, state, and local law enforcement in addressing threats to national security.¹⁶² The development of Fusion Centers was based on three shifts in post-9/11 policing:

(1) a broader adoption of proactive policing strategies in addition to those that were often viewed as largely reactive, (2) widespread recognition of the significance of state and local law enforcement in homeland security efforts, and

(3) a consensus regarding the importance of institutionalizing the sharing of information among law enforcement entities.¹⁶³

The federal government viewed local entities as critical to national counterterrorism activities, since, “for the most part, information about homegrown plots is available only through domestic policing activities.”¹⁶⁴

Alongside the creation of eighty Fusion Centers,¹⁶⁵ federal funding also went to local law enforcement agencies to implement surveillance technologies.¹⁶⁶ The Justice Assistance Grant (JAG) program of the Department of Justice, for example, awarded approximately $3.8 billion to state and local law enforcement between 2003 and 2014.¹⁶⁷ In 2012 alone, JAG funds were used by 113 state and local jurisdictions to purchase license plate readers, 371 jurisdictions to purchase video camera equipment, and 284 jurisdictions to purchase undercover surveillance equipment.¹⁶⁸

Applying the Fusion Center rubric of multi-agency integration, Real Time Crime Centers (RTCCs) were created to harness data for the aims of local law enforcement, shifting their approach towards prediction strategies. Beginning in the early 2000s, RTCCs emerged as the most coordinated local law enforcement surveillance systems, with over 147 nationwide.¹⁶⁹ Their goal, according to the Department of Justice, is to “centralize a broad range of current and evolving technologies, coordinate sworn and/or nonsworn human resources, and direct the attention of both to high-crime areas, active crimes in progress, large-scale public events that may require law enforcement presence or response, and/or high-profile or highly recidivistic offenders in the community.”¹⁷⁰

An RTCC can integrate and display data from numerous sources, including emergency/911 dispatch systems, local record management systems (to determine prior convictions and outstanding warrants), video cameras (police, government, and private), electronic monitoring systems (to track individuals on supervised release), license plate readers, facial recognition software, social media applications, geolocation data, and FBI databases.¹⁷¹ “Each RTCC is slightly different, but their function is the same: gather surveillance data across a city and use that to build a live picture of crime in the city.”¹⁷² A common and salient feature of RTCCs is the integration of networked surveillance cameras. Imagine a wall of video screens and multiple consoles, with civilian analysts and law enforcement attached to headphones, rapidly deploying information from a variety of sources to on-the-ground police. This is the idealized version of an RTCC, as depicted on surveillance industry websites.¹⁷³

The first RTCC was launched in New York City in 2005 at the cost of $11 million, largely funded by the New York City Police Foundation.¹⁷⁴ The center was staffed continuously by forty detectives and analysts, with a two-story wall of video screens and connections to police databases and mapping information, allowing access to over 33 billion public records, 31 million national crime records, and 25 million state criminal records and complaints “in a matter of minutes.”¹⁷⁵ The RTCC also incorporated access to an unknown number of public and private surveillance cameras.¹⁷⁶ A 2021 Amnesty International report, based on volunteer assessments of 45,000 NYC intersections, identified over 25,500 cameras in public, with an estimated 2,266 readily identifiable as New York Police Department or Department of Transportation equipment.¹⁷⁷ Initially, this RTCC focused on shootings and homicides, but it quickly expanded to cover all crimes.¹⁷⁸

New York City’s surveillance system became a model for other cities, such as Memphis, Tennessee.¹⁷⁹ The Memphis RTCC is equipped with forty-two large-screen monitors, gunshot recognition technology, access to multiple databases, license plate readers, and, starting in 2010, Skycop cameras.¹⁸⁰ Skycop is one of the most sophisticated law enforcement surveillance networks in the country. Each of its cameras can pan and zoom, as manually directed by an analyst at the city’s RTCC.¹⁸¹ The footage can then be searched and recorded for later review.¹⁸² From 2010 through 2023, Memphis deployed over 2,000 Skycop cameras around the city, “all of which send back live feeds to the Police Department’s real-time crime center.”¹⁸³ In the first eleven years, Memphis spent over $10 million on Skycop surveillance technology.¹⁸⁴ As recently as 2024, the Memphis Police Department (MPD) received almost $2 million in federal funding, half of which was earmarked for surveillance cameras as a “vital tool to help make up for [the] MPD’s lack of manpower.”¹⁸⁵

Around the country, federal funding for RTCCs continues to flow to local jurisdictions.¹⁸⁶ The American Recovery and Reinvestment Act of 2009 (Recovery Act) earmarked over $4 billion for state and local police.¹⁸⁷ At least sixteen jurisdictions have also used funds from the American Rescue Plan Act of 2021 (ARPA) to increase video surveillance, and four have used the federal money to expand or launch RTCCs.¹⁸⁸

ARPA grants have allowed mid-size and smaller jurisdictions to invest in surveillance technologies.¹⁸⁹ In Washington, $5 million in federal pandemic relief dollars was used to help fund Spokane County’s multi-city RTCC.¹⁹⁰ The system includes forty-six automatic license plate readers at high-traffic intersections, which cross-check with FBI national databases,¹⁹¹ and AI-powered video analytics software, which makes days of footage easily searchable.¹⁹² Similarly, other mid-size cities, like Mesa, Arizona and Richmond, Virginia, also received federal ARPA dollars for police surveillance purposes.¹⁹³

While vast amounts of federal funding for local law enforcement technology have spanned the War on Crime, the War on Terror, economic relief during the Great Recession (the Recovery Act), and COVID-19 stimulus aid (ARPA), measurable data on the effectiveness of this technology has largely absent and, when it does exist, lukewarm.¹⁹⁴ As noted by Professors Barry Friedman and Elizabeth G. Jansky, “Digital technologies are being used widely to monitor the populace, but once again, it is unclear whether these technologies are beneficial or problematic, particularly when one considers the social costs, which too often are ignored.”¹⁹⁵ This disinterest in measurable outcomes evinces a “technology for technology’s sake” mindset, which is characteristic of techno-solutionism. Not only has federal funding supported the spread of surveillance among local law enforcement by financing the purchase of new technology, it has also encouraged a techno-solutionist shift to ever more technologically intensive policing priorities, such as data gathering and crime prediction.

B. Corporate Partnership and Profit

The availability of vast amounts of federal funding has not escaped corporate interest. As stimulus and relief legislation like the Recovery Act and ARPA earmarked billions of dollars for local law enforcement, private technology companies jockeyed for government contracts.

A look at industry websites shows how explicitly surveillance companies have connected the dots between federal funding sources and corporate profits. Samsung, whose subsidiary, Hanwha Techwin, has cornered 16.7% of the North American security camera market share, states on its website, “Not surprisingly, RTCCs can be costly to implement. Fortunately, there is significant federal funding available that can be accessed to help cities and counties establish or significantly expand an RTCC.”¹⁹⁶ Similarly, Motorola operates a free grant assistance program, with grant consultants who “will do everything [they] can to assist you in submitting a successful grant application.”¹⁹⁷

Fusus, a Georgia public safety software company, has the largest market share of RTCCs, powering over 250 law enforcement jurisdictions.¹⁹⁸ Their goal, according to CEO Chris Lindenau, is to build “New York-style surveillance systems for thousands of smaller police departments . . . .”¹⁹⁹ One such smaller jurisdiction is Beaumont, Texas, with a population of approximately 115,000. The city entered a contract with Fusus in 2023 to set up an RTCC, at the cost of $275,000 annually.²⁰⁰ In May 2024, city police invited the Beaumont Independent School District to join the “growing list of partners that have Fusus cameras set up on site.”²⁰¹ If the partnership is approved, 1,500 cameras will be installed on school district property, costing $96,525 annually.²⁰² Overall, Fusus is estimated to have contracts with local law enforcement that range from $4 to $7 million annually,²⁰³ with individual contracts ranging from $50,000 to $150,000 a year.²⁰⁴ In addition to Fusus, Mortorola Solutions²⁰⁵ and Genetec²⁰⁶ are also major vendors of law enforcement RTCC capabilities.

Private companies providing policing tools is nothing new. What is new is the way that this “consumer-vendor relationship . . . poses greater concerns when the product itself is central to the development of the governmental suspicion that underlies so many enforcement decisions.”²⁰⁷ Surveillance companies have become more than vendors; they have become law enforcement’s “strategic partners, essential for advancing an organization’s technological capabilities,” according to Los Angeles Police Information Technology Bureau Deputy Chief John McMahon.²⁰⁸ “The relationships between surveillance technology vendors and police departments show the increasing degree to which private companies can guide, shape, and limit what the public police do.”²⁰⁹

One example of how private companies have inserted themselves into calculating suspicion is the development and deployment of facial recognition technology (FRT). Approximately 25% of police departments in the United States use FRT, with law enforcement facial recognition networks covering 117 million Americans.²¹⁰ Additionally, seven federal law enforcement agencies within the Departments of Homeland Security and Justice use FRT, including the FBI and Customs and Border Protection.²¹¹

FRT “automates face detection and identification by comparing captured images against digital photos from various sources, including surveillance footage, mug shots, and the internet.”²¹² The ability of FRT to accurately compare two images to determine whether they represent the same person is based upon software algorithms, most of which are proprietary and protected from public disclosure.²¹³

FRT algorithms have faced criticism for racial and gender bias.²¹⁴ These algorithms must be trained on high volumes of data, and “algorithms trained with biased data have resulted in algorithmic discrimination.”²¹⁵ In 2019, the U.S. Commerce Department studied 189 algorithms from ninety-nine developers, finding that the datasets used to teach FRT are overwhelmingly composed of lighter-skinned subjects.²¹⁶ The report found large false positives rates in West and East African and East Asian people, meaning that individuals from these backgrounds were more likely to be misidentified as a “match” by local law enforcement.²¹⁷ FRT misidentified women, people of color, and young people at high rates, with “darkerskinned females” suffering error rates of 34.7%.²¹⁸

News reports continue to document harmful examples of false identifications and other significant issues that result from the technology’s limitations.²¹⁹ Amazon’s Rekognition system was specifically criticized in a well-publicized M.I.T. report that found that it fared far worse than IBM and Microsoft’s FRT softwares.²²⁰ Despite these concerns, private companies, like Amazon, IBM, and Microsoft, began pitching FRT to law enforcement in late 2016.²²¹ Their potential windfalls were huge – the global market for FRT grew from approximately $3 billion in 2016 to an estimated $8 billion in 2022.²²²

In response to criticisms of Rekognition, Amazon initially defended its technology, before instituting a 2020 moratorium on law enforcement use of its software.²²³ However, in 2024, the Justice Department confirmed that the FBI was in the “‘initiation’ phase of using Amazon Rekognition” once again.²²⁴ Similarly, in the wake of the George Floyd protests in 2020, IBM announced that it would no longer sell FRT,²²⁵ only to sign a $69.8 million contract in 2024 to provide FRT to British immigration and law enforcement agencies.²²⁶

In the end, American companies seem willing to countenance the use of their products by law enforcement, even with the risks of racial and gender bias. This is unsurprising, as voluntary market constraint over the long-term would be incompatible with profit-oriented decision making. It is, however, consistent with Professor Shoshana Zuboff’s theory of surveillance capitalism, where “technology was, is, and always will be an expression of the economic objectives that direct it into action.”²²⁷ Here, the economic objectives motivating private surveillance companies are tied to their ability to increase sales to their customers: police departments. In a techno-solutionist mindset that values quick solutions over complex structural responses, neither law enforcement nor private companies have incentives to address the larger socio-economic systems or histories of racial and economic subordination that undergird the criminal legal system.

C. Participatory Surveillance

Direct partnerships with local police departments, although lucrative, are limited by the number of law enforcement agencies that are large enough to create and maintain RTCCs. One of the most potentially profitable markets for surveillance, however, is not law enforcement at all, but private constituents. Merging private and public cameras creates new profit sources, as well as a more thorough surveillance network. The result has been the expansion of “participatory surveillance” in policing, as private and public sources of information coalesce.²²⁸

While the encouragement of constituent participation in law enforcement surveillance is not new – the “If You See Something, Say Something” campaign was launched (and trademarked) immediately after the September 11, 2001 attacks²²⁹ – the ease of participation, the extent of cooperation, and the level of integration today is remarkable. This technologically facilitated deputization of private citizens for government surveillance creates several ruptures with traditional modes of surveillance. “A conventional understanding of surveillance is that it is a hierarchical system of power. This common understanding is represented in familiar metaphors such as ‘Big Brother’ and ‘Panopticon,’ both of which illustrate a vertical, hierarchical power relation between the gaze of the watcher that controls the watched.”²³⁰ The escalation of private surveillance complicates this linear hierarchy between watcher and watched. Surveillance is no longer strictly vertical – it can be a lateral practice as well.²³¹

1. Private Individual Cameras

Increasingly, RTCCs and other law enforcement surveillance systems “functionally blur the lines between private and public surveillance sources.”²³² The Detroit Police Department (DPD) pioneered “the first public-private-community partnership of its kind,” with Project Green Light, launched in January 2016.²³³ Using crime analysis techniques, the DPD determined that, “street segments characterized by gas stations, bars and liquor stores, and convenience stores, particularly in areas with illicit drug selling, were at disproportionate risk for violent crime, particularly shootings and robberies.”²³⁴ Based upon this data, the city created a formal partnership to connect private business cameras to law enforcement’s surveillance network.²³⁵

Initially, eight gas stations were equipped with real-time security cameras that fed directly to the city’s new RTCC.²³⁶ By 2024, over 900 cameras had been placed in local business and community locations.²³⁷ A detailed map on the City of Detroit’s website indicates the location of each private camera, each of which transmits directly to the city’s RTCC,²³⁸ where they are integrated with over 4,300 total cameras located throughout the city.²³⁹

Participating businesses purchase “high-definition surveillance cameras, lighting, and signage that identifies their location as a participant in the initiative,” along with an “external steady-burn green light visible from passing vehicles.”²⁴⁰ Installation and other set up costs add up to approximately $5,000, paid for by the private customer.²⁴¹ The DPD’s website lists options for leasing or purchasing cameras from a list of eight private vendors, including Comcast, Cronus, and Guardian Alarm.²⁴² The least expensive option, via a lease with Cronus, costs $449.95 for installation and $179.95 per month.²⁴³

Similar private surveillance integration can be found across the country. The Electronic Frontier Foundation counts 529 police departments that have launched private security camera registries, and this number only keeps growing.²⁴⁴ In early 2024, for example, the Memphis Police Department launched a camera registry called Connect 2 Memphis.²⁴⁵ The program gives constituents and businesses two options: 1) to either register their private surveillance cameras with the police department, allowing the MPD to contact the individual if they believe a registered camera might be useful to an investigation, or 2) to fully integrate their camera with law enforcement systems.²⁴⁶ Full integration “take[s] community security one step further” by allowing the MPD to access the private camera directly, without case-by-case authorization, integrating the private video feed directly into the MPD’s monitoring system.²⁴⁷

One of the biggest private-police surveillance networks is maintained by the Atlanta Police Department. Atlanta is the most surveilled city in the United States, with almost fifty cameras per 1,000 people, and “the seventh most surveilled city in the world.”²⁴⁸ As of 2021, there were 24,000 surveillance cameras in Atlanta, a city with a population of approximately 500,000.²⁴⁹ In early 2022, Atlanta launched Connect Atlanta which, like Detroit and Memphis, integrates private surveillance into law enforcement systems.²⁵⁰ As of February 2025, Connect Atlanta included over 22,460 registered private cameras and 20,313 private cameras fully integrated into Atlanta’s RTCC, giving police access to over 42,000 private cameras.²⁵¹

2. Business and Neighborhood Associations

Registries of individual constituent cameras are not the only emerging model of participatory surveillance. Business interests and neighborhood associations have also coordinated their surveillance efforts, both with each other and with law enforcement. San Francisco, for example, has approximately sixteen Community Benefit Districts (CBDs), which are neighborhood organizations formed to provide services to the area.²⁵² One of the services is camera surveillance. Starting in 2012, billionaire “tech-mogul” Chris Larsen began funding the installation of video surveillance cameras and the CBDs that monitor them.²⁵³ By 2020, Mr. Larsen had invested $4 million to create a network of over 1,000 cameras, covering 135 city blocks.²⁵⁴ One of the most well funded CBDs is the Union Square Business Improvement District, which operates approximately 400 high-definition cameras in the city’s downtown retail district.²⁵⁵ The cameras are sophisticated and are capable of license plate reading and facial recognition (although these capabilities are not known to have been utilized yet).²⁵⁶

From Mr. Larsen’s perspective, “The community groups can get this done faster than if you were to make this a citywide government-sponsored project which would cost a ton of money . . . .”²⁵⁷ This view is echoed by the members of the CBDs: “If you went to the board of supervisors and asked the members to approve this, you’d end up having a conversation about government and surveillance.”²⁵⁸ These comments, however, point to a concern with privately funded and operated surveillance systems – they operate beyond the reach of public oversight and laws aimed at government surveillance.

In San Francisco, the use of technological surveillance by law enforcement in the late 20th and early 21st century has been subject to hard-fought, democratically pursued safeguards.²⁵⁹ For example, the San Francisco Police Commission instituted Department General Order 8.10 in 1990, one of the first local rules to place limits on police surveillance.²⁶⁰ The order requires “articulable and reasonable suspicion” of criminal activity for SFPD investigation of civilian activities that implicate the First Amendment, such as demonstrations, rallies, and marches.²⁶¹ In 2019, San Francisco became the first city in the nation to ban police use of FRT, with its Stop Secret Surveillance Ordinance.²⁶² In 2022, the SFPD Board of Supervisors imposed a policy that limited continuous live-feed monitoring by police to twenty-four hours after access is provided.²⁶³

Privately funded surveillance, like the kind operated by the CBDs, can operate outside the reach of these types of local government safeguards. While the 2019 Stop Secret Surveillance Ordinance prohibits the use of FRT by law enforcement, it does not ban its use by private operators.²⁶⁴ Under the 2022 policy, private cameras can still live-feed indefinitely.²⁶⁵ According to Dave Maass, a senior investigative researcher at the Electronic Frontier Foundation, “San Francisco has passed a very sophisticated surveillance ordinance that bans facial recognition by the Police Department, but yet you have these independent agencies within the city limits making their own decisions.”²⁶⁶

This has created loopholes for local law enforcement, who have accessed private surveillance systems, despite the prohibitions in place. For example, in the wake of George Floyd’s death, “[f]rom May 31 through June 7, 2020 . . . the San Francisco Police Department (“SFPD”) acquired, borrowed, and used a private network of more than 400 surveillance cameras to spy on protestors in real time.”²⁶⁷ Essentially, privately controlled cameras became a work-around to democratically implemented restrictions on police surveillance. Moreover, as will be discussed in Part III.A, merging private and public cameras creates a surveillance network distorted by the demographics of who has the affluence and actionable desire for surveillance technology. The results have been patterns of racially biased surveillance.

The history of law enforcement CCTV, as described above, offers a view into the factors which played decisive roles in the techno-solutionist orientation of police surveillance. Federal funding, with roots in the racialized politics of 1960s War on Crime strategies, provided billions of dollars in grants to local enforcement for increasingly militarized policing infrastructure, including surveillance.²⁶⁸ Post-9/11, federal funding re-oriented local law enforcement towards strategies of data integration and crime prediction, helping generate RTCCs around the country.²⁶⁹ Corporate interests followed, which led to the profit-motivated spread of police surveillance to midsize and smaller jurisdictions, despite concerns about the racially biased repercussions of algorithms and technologies.²⁷⁰ Finally, participatory surveillance has deputized private individuals and businesses to expand the reach of law enforcement and create a more thorough surveillance network.

From this history of law enforcement CCTV, there emerges a techno-solutionist approach to problem-solving that has fundamentally reoriented 21st century policing. The emphasis has been on prioritizing (and profiting) from one “big idea” – surveillance technology – as a magic bullet, without equally focusing on local contexts, structures of power, or systemic inequities. The juggernaut of policing surveillance moves forward, despite its exacerbation of racialized harms and the absence of clear public safety results. The simple, linear narrative framing technology as a tool to be applied to existing, predetermined problems that are easily quantified and ahistorical does not match the reality of how technology has changed policing goals over the last fifty years. Technology has become an end in and of itself.

III. Racialized Surveillance and the Suspect Data Double

Real-time video surveillance is one of many technological tools used by 21st century police departments. Like other such tools – body cameras, license plate readers, GPS monitoring, and cell phone tracking, for example – it works within the existing racial inequities of the criminal legal system.²⁷¹ The insertion of technology into policing does not remove this stain.²⁷² “Technology is neither good nor bad; nor is it neutral.”²⁷³ The growth of police surveillance reflects the same unjust racial biases inherent in the criminal justice system overall. Police surveillance is deeply connected to the broader history of policing Black and Brown communities and cannot be separated from these origins. “For the communities who bear the brunt of policing and the other facets of the criminal system, law enforcement technologies only reify and exacerbate injustice.”²⁷⁴

This critique of techno-solutionism cannot be addressed by simply adjusting algorithms. “[T]he harms that accompany the deployment of new technologies in context are not simply the result of bias, but betray a fundamental mismatch between complex social issues and tech solutionism.”²⁷⁵ Techno-solutionism reifies the apolitical quick-fix, ignoring consequences that extend beyond the immediate short-term problem. As argued by Simone Browne, tech-driven police surveillance is the latest chapter in the long American history of policing Black communities.²⁷⁶

Video surveillance has continued to result in racially biased monitoring. The sources of these biases can be divided into several categories. This Part focuses on two such areas: first, on the location and ownership of the cameras themselves; and second, on the implications of pervasive surveillance and Big Data on racialized policing.

A. Racially Patterned Surveillance

One factor that can create large, racialized differentials in terms of who is subject to surveillance is the location of the camera itself. On a macro, multi-jurisdictional level, video surveillance has been adopted more pervasively by larger jurisdictions. While almost nine out of ten jurisdictions with populations greater than 250,000 operate video surveillance systems, that ratio drops to 49% when including all local law enforcement.²⁷⁷ Larger police departments prioritize strategies such as targeting high-risk areas, implementing saturation patrols, and identifying repeat offenders – activities that require “more advanced information systems and data analytics” – at a higher rate than smaller jurisdictions.²⁷⁸ A 2020 survey of local police departments found that “larger jurisdictions tend to have the highest concentration of police departments with access to surveillance technologies . . . ,” with the exception of body cameras.²⁷⁹ Jurisdiction size is not only correlated with higher levels of surveillance, but also with racial diversity. The result is that “jurisdictions with the greatest proportion of non-Hispanic white residents are . . . less surveilled by every technology” when compared to jurisdictions with the smallest proportion of non-Hispanic white residents.²⁸⁰

This disproportionality also intensifies within any given jurisdiction. Despite San Francisco’s policy against monitoring based in any part on race,²⁸¹ looking at law enforcement’s network of cameras reveals oversurveillance of Black and Brown neighborhoods. Through a public records request, the Electronic Frontier Foundation obtained the locations of 2,753 public and private cameras linked or registered with law enforcement.²⁸² The map shows huge swaths of the city escaping any police video surveillance, including Pacific Heights, the Sunset, Richmond, Haight-Ashbury, Noe Valley, West Portal, and Glen Park – all regions with high proportions of white residents.²⁸³ The vast majority of the surveillance is situated in areas with higher proportions of Black and Latino residents – the Bayview, Mission, Western Addition, the Tenderloin, and SoMa.²⁸⁴ The other areas with large clusters of cameras are in commercial or tourist zones – Union Square, Chinatown/ North Beach, Fishman’s Wharf, and the Embarcadero.²⁸⁵

This dual distribution pattern – high surveillance in areas of commerce and tourism, as well as in communities of color – is echoed in New York City. Amnesty International’s 2021 count of public-facing cameras shows correlations between race and surveillance levels. In the Bronx, Brooklyn, and Queens, higher percentages of nonwhite residents were correlated with higher levels of surveillance.²⁸⁶ The high levels of surveillance in Manhattan are not correlated with higher percentages of nonwhite residents but with higher commercial, transportation, and tourism activity.²⁸⁷ New York City, as a whole, shares a similar dual-distribution pattern to the one found in San Francisco.²⁸⁸

The dual positioning of surveillance cameras in Black and Brown communities and business districts, however, is not a consistent practice across the country. A second pattern emerges as well, which intersects with participatory surveillance and property interests in telling ways. In St. Louis, Missouri, for example, an investigation uncovered that “there are significantly more cameras in affluent, majority-white areas than there are in majority-black, impoverished parts of the city.”²⁸⁹ Although law enforcement has emphasized the violent crime fighting benefits of video surveillance, there are few cameras in North St. Louis, where a majority of the city’s homicides have occurred.²⁹⁰ North St. Louis is 90% Black and, in certain areas, lacks an underground fiber network to facilitate real-time video feeds.²⁹¹ Wealthy (and white) Central West End, in contrast, has a much larger number of video cameras, disproportionate to both its population and crime rate.²⁹² Downtown St. Louis has the highest concentration of city-owned cameras and a relatively lower homicide rate than North St. Louis.²⁹³

One reason for this disparity is that the vast majority of the cameras, which run the expensive Genetec platform, are privately paid for, either by private businesses or business districts.²⁹⁴ Those that are city-owned are paid for by taxing the specific district where the camera will be installed. St. Louis’ RTCC, fed mainly by the participation of civilians and business organizations that can afford cameras, is effectively “a taxpayer-funded security monitoring service for parts of the city that can afford cameras.”²⁹⁵

Memphis’s experience with Skycop is similar. The initial placement of Skycop cameras, starting in 2010, was, by and large, out of the police department’s hands. Of over 2000 cameras, roughly 1,300 were “bought by community groups and businesses that install[ed] camera boxes or mobile camera units at perceived trouble spots.”²⁹⁶ These constituents, mostly in affluent neighborhoods, then donated the cameras to the police department for continuous, real-time uploading at the RTCC. The Memphis Downtown Neighborhood Association, for example, raised more than $30,000 to install four Skycop cameras in the immediate area.²⁹⁷ Each camera cost between $6,000 to $8,000.²⁹⁸

For years, Memphis has accepted these camera donations and agreed to place them where the donating constituents wanted them: “guarding either the city’s valuable Downtown developments or its wealthiest East Memphis neighborhoods.”²⁹⁹ Regardless of who purchased the cameras, “taxpayer dollars cover[ed] the cameras’ . . . annual cost of about $42,000 a year . . . .”³⁰⁰ An investigation uncovered that the areas with the most crime had disproportionately fewer cameras.³⁰¹

These two patterns – the placement of publicly funded surveillance cameras in Black and Brown neighborhoods versus the placement of self-funded cameras in wealthy neighborhoods and business districts – point to the ways that camera placement can be racially biased in multiple directions. In cities such as San Francisco, law enforcement surveillance was initially publically funded and cameras were placed disproportionately in communties of color. In other cities, such as St. Louis and Memphis, initial camera surveillance was privately financed and placed in relation to property interests.

While this can be seen as alternative patterns of surveillance, it can also be seen as the same surveillance trajectory, just with different initial positions. The real distinction between San Francisco and St. Louis’ versions of video surveillance might simply be the original funding sources, which were also affected by timelines of technological development. San Francisco’s first Community Safety Cameras were installed in 2006,³⁰² prior to the widespread availability of participatory surveillance technology.³⁰³ St. Louis launched its RTCC almost a decade later, in 2015, when commercially available surveillance tools such the Ring doorbell camera were available.³⁰⁴ Those jurisdictions that depended highly on participatory surveillance and constituent funding of law enforcement surveillance at the outset moved next to expanded surveillance, starting with high-crime areas and communities of color. By contrast, those jurisdictions which began with publicly funded surveillance developed in areas of commerce and high-crime areas first, before incorporating private and participatory surveillance mechanisms. Moreover, the lines between public and private surveillance have blurred, as private participatory surveillance has become increasingly integrated into law enforcement networks. The resulting trajectory leans towards pervasive surveillance.

B. Pervasive Surveillance and Big Data

The move from specific foci of police surveillance to more generalized, constant surveillance has occurred alongside two critical developments in surveillance systems. First, surveillance has expanded into the private home, incorporated into all types of smart home devices, ranging from nanny cams to smart doorbells.³⁰⁵ Second, public surveillance systems have been increasingly integrated with artificial intelligence capabilities, utilizing features such as FRT and license plate readers.³⁰⁶

One significant outcome of these twin trends is the ascendance of Big Data.³⁰⁷ It has been applied to everything from the best time to purchase a plane ticket to the prediction of epidemic outbreaks.³⁰⁸ The hallmarks of Big Data, as touted on industry websites, are 1) increased volume (the sheer amount of data points), 2) greater variety (a heterogenous range of sources and data types, including unstructured data, such as text and audio, that require additional processing), and 3) higher velocity (the rate at which data is received, including at or near real-time, relative to data production) – the three “Vs.”³⁰⁹

The power of Big Data lies in its ability to combine diverse types and sources of information and derive correlations and probabilities. In the context of policing, the data that is combined to create a fuller picture of an individual’s activities might not originate with law enforcement or have any connection at all with crime. It need not be limited to data points like criminal warrant records, stay-away protective orders, or license plate registrations. The purposes of the data gatherers might be disparate and even oppositional.³¹⁰

Big Data has ushered in a new relationality between different types of information. “Big Data is fundamentally networked. Its value comes from the patterns that can be derived by making connections between pieces of data, about an individual, about individuals in relation to others, about groups of people, or simply about the structure of information itself.”³¹¹ This capability renders the search for theories connecting data – theories of causation – unnecessary. With massive amounts of data, the traditional scientific method of “hypothesize, model, test . . . is becoming obsolete.”³¹² The why people do what they do is not as significant as the very fact that they do what they do. In some cases, their actions can be predicted based on the analysis of huge amounts of data, not models or theories.

Big Data, like much of technology, enjoys a status as “objective” or “less biased.”³¹³ However, as argued by sociologist Sarah Brayne, the algorithms that sort and connect the pieces that make Big Data so powerful

do not transcend the social, but are shaped by the social world in which they are created and used. The activities that generate and the technologies that analyze data are all embedded within social contexts and power structures, so the resulting data are anything but “natural,” detached, or purely descriptive.³¹⁴

In the criminal legal context, this “hubris” can be noted in the first wave of algorithmic interventions that assessed sex offender recidivism risk³¹⁵ and Public Safety Assessments that determined whether to grant pretrial release. These algorithmic risk assessments have been criticized as racially biased.³¹⁶ The inputs into these algorithms – prior arrests, convictions, prison sentences, failures to appear in court – are drawn from police systems that disproportionately affect Black and Brown individuals.³¹⁷ This leads to racially biased outcomes, resulting from the simplicity of the algorithms, since they have a quantifiable number of inputs that are almost exclusively derived from law enforcement sources. The disparate impact of algorithms is not unique to the criminal legal system, as scholars have shown their effects on decisions related to public benefits, housing and financial markets.³¹⁸

How does Big Data thinking intersect with the police surveillance apparatus? As jurisdictions move towards pervasive surveillance, policing strategy and the purposes of police surveillance have changed.³¹⁹ Deterrence and crime-apprehension goals have given way to data gathering and analysis. Surveillance itself has transitioned from the “familiar ‘legacy’ version of targeted, purposeful spying” to “the emerging model of ubiquitous, opportunistic data capture.”³²⁰ The result is a more generalized form of surveillance, unmoored from any individualized suspicion of those being watched.³²¹ It is the difference between a single police video camera installed across the street from a suspected methamphetamine distributor³²² and a real-time ceaseless video feed placed at every intersection in every neighborhood. The primary goal of police surveillance is no longer to deter crime or to even capture evidence of a crime on video. The goal is the accumulation of data, which can then be saved, sorted, searched, and combined with FRT, automated license-plate readers, cell-phone tracking, drone camera feeds, and/or biometric data from GPS monitoring.³²³ In the hands of law enforcement, this strategy of surveillance has led to the creation of what I call the “suspect data double.”

The suspect data double imports the concept of the “data double” from sociology and surveillance studies and applies it to suspects in the criminal legal system.³²⁴ Generally, the data double is a profile constructed from aggregated data where any data point of the profile may be connected to any other.³²⁵ It describes “the representations and informational flows generated by the users of networked communications technology to allow the agencies (corporate; governmental; and supra-governmental) managing those systems to construct and monitor models of [their] users.”³²⁶ With the growth of Big Data, the data double’s contours continue to expand, including government records, financial transactions, internet searches, cell-phone usage, location information, etc.³²⁷

The data double is most often used in marketing or targeted political campaigns.³²⁸ For example, Google and Facebook use biometric information such as age, gender, location, browsing history, purchasing patterns, likes, social media posts, etc. to create data doubles to identify which specific users to target with advertisements.³²⁹ With Big Data surveillance moving into policing, the suspect data double has become the central character in predictive policing. Predictive policing refers to a set of law enforcement practices that includes the “proactive use of algorithmically mediated data analysis for the purpose of finding patterns in datasets, based on which risk estimates are produced for either individuals or locations . . . .”³³⁰ These estimates can then be utilized by police for pre-emptive action, targeted deployment or other strategic interventions.

Predictive policing targets suspect data doubles, not suspects. It casts a wider net than previous police determinations of suspicion, since everything can be evidence of suspicion if it correlates with other data in specific ways. The suspect data double is the culmination of the vast amounts of information harvested from ubiquitous surveillance and the evolution of surveillance as a tool of crime prediction and data aggregation. The potential implications of the suspect data double will be the focus of future scholarship, specifically the ways in which the use of rhizomatic data profiles and the growth of the preventive state will continue to enable law enforcement’s disparate impact on communities of color.

Conclusion

In 2021, the Seventh Circuit described a panoptic future, in which “Americans will traverse their communities under the perpetual gaze of . . . millions of unblinking eyes . . . ,” cataloging their work routines, their recreational activities, their personal associations, and even their attire, belongings, and snippets of their conversations.³³¹ Even as it declined to apply Fourth Amendment protections to constant, warrantless, and prolonged police camera surveillance, the Court acknowledged that its decision was a “harbinger” of the challenge of applying existing Fourth Amendment jurisprudence to rapidly evolving surveillance technologies.³³²

By 2025, this foretold future has unquestionably arrived, and with it, new modes of policing. Techno-solutionism has recruited surveillance technologies to address police staffing challenges, and in doing so, has written the newest chapter on how technology is changing policing goals. While traditional policing focused on deterrence and apprehension, the pervasive surveillance machines of the 21st century have ushered in a shift towards crime prediction and prevention.

Critically, these techno-solutions have flowed to points beyond the reach of the Fourth Amendment and most Constitutional limits.³³³ As Professor Carol Steiker points out, while “[o]ur federal Constitution has a lot to say about crime and punishment,” it has less to say about the “limits of the state not as punisher . . . but rather as preventer of crime and disorder generally.”³³⁴ The limitations on the preventive state pale in comparison to those on the state as punisher, partly due to the historical context of our Constitution. While “the dangers of the punitive state were well known” to the Constitutional drafters, “[t]he preventive state became possible only as the next century progressed”³³⁵ with technological advancement.

This Article examines how technology, far from a passive tool, has been an agent in altering the goals of policing towards predictive action. Technology has established a new mode of policing that is cloaked in seemingly neutral techno-solutionism and is positioned to weather the staffing shortages and the socio-political reckoning faced by law enforcement in the 21st century. The ideology of techno-solutionism, in ignoring the social and historical complexity of human problems, will continue to extend the preventive state and amplify the entrenched racial biases of our criminal legal system.