In the spring of 2013, the Oakland City Council was on track to approve one of the first and most extensive systems of public surveillance in the world. Intending to police the Port of Oakland—one of the largest ports in the U.S. and plagued by a record number of crimes—the Domain Awareness Center (DAC) would be quickly expanded across the entire city. The system combined a wide network of surveillance cameras and gunshot and thermal intrusion detectors into a data stream for police monitoring. Cameras would be placed all over Oakland, including public schools and the Coliseum complex. City officials saw DAC as a way to make their policing smarter and more efficient. Emergency responses would improve, a benefit especially helpful in the event of a mass shooting or terrorist attack. During its development, however, few citizens were aware of the proposed program. After the Snowden revelations broke, public skepticism of government surveillance heightened and the project faced immense opposition from privacy and civil rights advocates. Eventually, the proposal was amended and the system now only monitors the Port of Oakland.
While this attempt at digitized policing by the Oakland City Council was unsuccessful, efforts toward greater public surveillance are taking place, often without our full awareness. Other cities in the U.S., such as New York, have moved forward and implemented analogous tools, integrating public surveillance technology with private CCTV networks and public transportation cameras. With government support, private companies are investing heavily in this market, projected to almost double in size, from 45 to 75 billion U.S. dollars, in the next five years. Yet, this investment flies in the face of scientific evidence that shows that video surveillance has very little effect on crime deterrence. A 2005 study for the British Home Office, the ministerial department responsible for national security in the U.K., another heavily surveilled country, concluded that “for the most part CCTV did not produce reductions in crime, and it did not make people feel safer.”1 At the same time, new investments in mass surveillance curtail citizen’s privacy and potentially divert funds from other initiatives like, for instance, hiring new police officers. If data-driven policing is the future of public safety, are these technologies really worth it?
In We See It All, Liberty and Justice in the Age of Perpetual Surveillance, Jon Fasman, the U.S. digital editor of the Economist, provides a comprehensive account of modern surveillance technologies adopted by American law enforcement. Fasman, an experienced reporter and detailed storyteller, has written extensively about the criminal justice system, underlining its inconsistencies and blind spots. In this book, Fasman comes off as neither a Luddite nor a techno-utopian. Writing in a space where the subject matter can become complicated quickly, he develops his argument with caution and has a penchant for poignant examples that unmask the ethical and political trade-off of surveillance. The book argues that at the moment, most surveillance technologies are legal. Yet, their legality mostly derives from the fact that they are so new that they operate in a regulatory gray zone. The legal void deprives citizens of the means to dispute their implementation and the ability to demand greater privacy. And, as Fasman reports, our privacy is often curtailed to very large extents and not always justified by ongoing investigations.
The current level of surveillance worldwide is unprecedented. In 2019, a report by IHS Markit showed that, in the U.S., there are more than 70 million security cameras, approximately one every 4.6 individuals.2 In comparison, China—a country commonly used as the epitome of a surveillance state—has one security camera for every 4.1 individuals. In the U.S., most footage is collected by private surveillance companies which often place cameras in places where people expect some degree of privacy, such as hotels, restaurants, and event centers. The privatization of surveillance also brings a new challenge: the interests of private companies often conflict with those of the surveilled. A surveillance company may, for instance, reduce the price of camera installation in exchange for a business’s permission to use their camera footage for research and development. Businesses may get a better bargain, but consumers pay the difference in reduced privacy.
Private sector involvement in law enforcement also transforms the nature of policing. Marketed directly to police departments, surveillance technologies often promise to amplify and extend existing policing capabilities. Overwhelmed agencies short on budget, may adopt them in the hope of better serving their citizens. Yet, as Fasman describes extensively, implementing technological solutions doesn’t just expand policing capabilities, they are an entirely new form of policing. In the pre-digital era, police officers did not record their surroundings, but police camera footage surveils a whole city and can be retained indefinitely.
Further, Fasman highlights that integrating surveillance cameras with facial recognition technology creates surveillance systems with unprecedented scope. Such a system could be helpful to track down criminals, but what about everybody else? A similar argument could be used for automatic license plate readers (ALPRs). Used by police and private individuals alike, ALPRs are small, rectangular boxes that can be stationed anywhere, most often found on car roofs and trunks. These devices automatically read license plates, collect plate information and enter this data into massive databases, including data such as the location of the car at the time of the reading. Such databases may prove helpful to recover stolen vehicles or track those suspected of illegal activities. But ALPRs log every plate they see, and 99.99% of cars are not involved in illegal activities, Fasman reports.
The astounding discrepancy between the needs and the capabilities of the surveillance system is a constant focus of the book, but such an unprecedented level of monitoring is hard to justify for several other reasons.
Mostly operated by small companies, such surveillance systems store information in databases with dubious security. Earlier this year, a group of hackers gained access to a database of a private security company containing 150,000 live feeds from cameras inside schools, hospitals, gyms, police stations, offices, and women’s clinics.3 Privacy violations like these do not only expose us to unwanted attention and embarrassment, they make us easier targets of criminal activities. Imagine a terrorist trying to plan a new attack. With access to so much information, he could study the best moment to strike, using surveillance cameras to determine how best to elude manned security. In an effort to cultivate a perceived sense of safety, we might have paradoxically made ourselves more vulnerable.
The unchecked adoption of surveillance technologies has also a hidden, more pernicious danger. The psychologist Daniel Kahneman in his book, Thinking, Fast and Slow, noticed how, when faced with a difficult question, we may swap that question for an easier one, often without noticing it. Several of the technologies reported by Fasman seem to result from such a reformulation—they are simplistic solutions to larger, intractable problems.
Take the problem of urban crime, for instance. Shotspotter is an audio technology that helps police enforcement detect gunshots. Fasman finds that, historically, only 7% of gunshots are reported to the police, either because people don’t recognize them, or because they imagine someone else would do it. Even when gunshots are reported, people often provide imprecise locations making their directions effectively useless. Shotspotter detectors are paired with an AI that is able, with surprising precision, to identify gunshots immediately and provide a location. Oftentimes, such speed helps save lives, but their presence only responds to violent crime after it has occurred. It does nothing to deter violent crime and is powerless against mass shootings. According to a 2019 report from Pew Research Center, handguns are responsible for 64% of gun murders in the U.S.4 If a city government has the goal of preventing violent crime before it happens, I believe gun control to be a more promising avenue than Shotspotter.
Another example of misplaced hope in technology has been the widespread adoption of body cameras. These devices are often praised for offering objective proof of police misconduct in litigation and seem to keep police officers in check, but several studies have claimed that they are not effective at preventing police violence. Investing massive resources to expand the use of body cams would prove a waste in absence of broader police reform that reinforces ideals of mutual respect and coöperation and establishes mutual trust between citizens and law enforcement.
Perhaps, caught up in the belief that technology is always useful, we forgot to look at what really works. At the end of the 1990s, researchers in economics, political science, and sociology were puzzled by the sharp drop in crime that characterized the previous decade. At the national level, homicides had declined by 43%, robberies 45.8%, and aggravated assaults by 26.7%.5 Such decline was unusual in the Western world. By comparison, between 1994 and 1999 the homicide rate in Europe declined only 4%. While some public enforcement reforms have been taking place, such a dramatic change was unexpected. In 1995, the Attorney General’s office commissioned a report to James Alan Fox, one of the most prominent criminologists, that projected that even under the most optimistic assumptions “the next crime wave will get so bad that will make 1995 look like the good old days.” When this prophecy didn’t materialize, commentators praised the effectiveness of new policing strategies, like those espoused by the mayor of New York, Rudy Giuliani.
They couldn’t have been more wrong. Study after study showed that the most significant factors inducing the drop in crime had very little to do with new approaches to policing.6 The drop in crime was due, in large part, to a wave of new hirings of police officers. More cops in the streets meant that criminals were effectively deterred and incidents could be quickly addressed.
At the end of his book, Fasman appeals to the readers to imitate the citizens of Oakland and demand that surveillance tools are regulated and that privacy protections put in place. We should also ask our legislators to look at the scientific evidence and question whether surveillance is really what we need to ensure public safety. It might open the door to even more crime.