It’s Not Too Late to Get Body Cameras Right

Two technology researchers argue that police-worn cameras have a long way to go before they're a tool for good.

Eric Miller / Reuters / Shutterstock / The Atlantic

Police-worn body cameras are coming. Support for them comes from stakeholders who often take opposing views. Law enforcement wants them, many politicians are pushing for them, and communities that already have a strong police presence in their neighborhoods are demanding that the police get cameras now. Civil-rights groups are advocating for them. The White House is funding them. The public is in favor of them. The collective—albeit, not universal—sentiment is that body cameras are a necessary and important solution to the rising concerns about fatal encounters between police and black men.

As researchers who have spent the last few months analyzing what is known about body cams, we understand the reasons for this consensus, but we’re nervous that there will be unexpected and undesirable outcomes. On one hand, we’re worried that these expensive technologies will do little to curb systemic abuse. But what really scares us is the possibility that they may magnify injustice rather than help eradicate it. We support safeguards being put in place. But the cameras are not a proven technology, and we’re worried that too much is hinging on them being a silver bullet to a very serious problem. Our concerns stem from three major issues:

  1. Technology doesn’t produce accountability.

  2. Removing discretion often backfires.

  3. Surveillance carries significant, hidden economic and social costs.

We’re concerned that in the rush to support the communities who face oppressive policing practices, during the pivotal moment when these issues are at the center of a national conversation, the problems of body cams are being glossed over. Even the researchers behind the most oft-cited study demonstrating the efficacy of body cams are skeptical about the generalizability of its results or the net benefits to community policing. They emphasize that more empirical research is needed to determine what the impact of body cameras are to policing.

Most importantly, we need assessment, research-driven implementations, and mechanisms that will allow us to stop going down this path if it is not enabling accountability or improving the well-being of communities of color.

* * *

Many communities rightfully want police and governments to be held accountable for brutal encounters between men of color and the officers who police them. Because the police-brutality incidents that have been caught on camera prompted widespread protests and community engagement, advocates hope that the mandatory collection of on-the-job footage will provide the evidence of police brutality that these communities need to prove their experience to the wider world—evidence that has often been missing in the past, and that is necessary to galvanize action to curb abuses or simply enable prosecution. Politicians and advocates hope that the mere existence of cameras may drive police officers to think twice before using force; in any case, advocates hope that camera evidence will make it easier to find and punish abuses that do occur. Journalists and the public simply want more transparency about policing practices. To get there, they want material evidence.

The police force also wants accountability. Officers want the video to exonerate them from false complaints. Cities want to reduce how much they pay out in lawsuits to victims of police misconduct (New York City, which is self-insured, paid out nearly $1 billion in the last decade). Prosecutors are leery of the cameras because they think they’ll be stuck reviewing hours of new footage with each case, particularly to make sure they don’t withhold any exculpatory evidence from the defense. At the same time, they’ve seen the way that dash-cam footage can help them prosecute cases successfully.

The temptation of technology as an accountability tool is not new, but accountability is not done by technology. Accountability is achieved by people and systems using tools like technology as part of their bureaucratic processes. There is effectively a global consensus that body cameras are a good thing to have because everyone has a different idea of what they’re agreeing to, a different model of appropriate bureaucracy. The bureaucratic and political battles over policies of use, access, and retention are not yet resolved, and they are significant. Who gets to see the footage, and in what circumstances, will matter. The features and capabilities of the technology matter. What happens when the camera reveals more about what was in the officer’s scope than what they could physically see at the time, especially at night? Or when cameras get additional features, like heat sensors? Even on basic practical questions, such as whether and when officers or the public should see the footage, there is no consensus.

Everyone is imagining what should be done when a citizen is shot dead by a member of the police, but few are focused on how footage might impact accountability or shift power dynamics in more routine encounters. For example, prosecutors might rely on the existence of footage, and a suspect’s ignorance of exactly what is captured, to obtain plea bargains more efficiently. Already, countless people agree to a plea bargain for crimes they didn’t commit because they don’t feel confident that they can prove their innocence, regardless of how much our system is supposedly about proving guilt. What happens when they are threatened with video footage or facing video footage that may be cropped, presented out of context, or otherwise manipulated to make them look more culpable than they really are? Who is empowered or disempowered by body cameras and their footage depends very much on the bureaucratic goals of the people making decisions about their use.

Already, officers are crediting body cameras with the ability to capture evidence in cases that are traditionally harder to prosecute because victims don’t want to press charges, such as domestic violence situations. Many victims of domestic abuse don’t want their partners jailed because of economic or familial considerations. In fact, policies that mandate arrests of domestic-violence offenders may intimidate victims from calling for police intervention in the first place—and those mandatory arrests may have a statistically minimal impact on future violence, while the intervention itself can have a significant deterrent effect. When victims decline to testify, officers and prosecutors see body cams as tools to provide the evidence they need to prosecute the perpetrator anyway, but that response may not be the best one for the victim’s situation. And when footage becomes available to either the public or the perpetrators, the potential consequences for the victim may outweigh the benefits of engaging law enforcement in the first place.

The availability of photographic and video evidence has helped bring the brutal encounters between communities of color and police to the center stage of national conversations. Rodney King. Michael Brown. Eric Garner. Walter Scott. However, footage has not necessarily produced arrests and prosecutions. It’s also not clear that body-cam footage will serve the same galvanizing force as citizen video. If anything, the “civilizing effect” of a body camera stems from an awareness of being watched.

Broader reforms to policing powers and the systems that hold them accountable are needed to move beyond the limited accountability that heightened cognitive awareness of surveillance can serve. Body cameras won’t alter legal doctrines and policing practices around the issue of “qualified immunity” for officers: Officers can use force, and even deadly force if they perceive, from the perspective of an officer in the heat of the moment, that their life is threatened or that the community is in danger when a suspect runs away. A camera won’t stop an officer from claiming that a suspect reached for something that they could reasonably guess was a gun, or from arguing that the subject was aggressive and threatening. Even when the footage suggests otherwise, officers are routinely given the benefit of the doubt in legal proceedings.

Research has also shown that we project our own biases onto the interpretation of video. What is “objective” is more complicated than many people realize, raising huge questions about the implications of video footage in legal proceedings. The Supreme Court uploaded dash-cam video in Scott v. Harris to its website as purportedly self-evident proof of a police officer’s reasonableness in that case. But when the Yale Law professor Dan Kahan showed the video to 1,350 people, they had varying interpretations of what happened that stratified notably by race, class, and political view. How people interpret video footage has more to do with their experience and attitudes towards the context than with any supposedly neutral analysis.

Too much is resting on the notion that body-cam video will speak for itself, but how footage is understood by the public and in court is not always clear. There is little doubt that citizen video has been valuable in raising questions about policing norms, but this doesn’t argue that police captured footage will be nearly as effective for the public, let alone in producing justice.

* * *

In the 1980s, many civil-rights leaders were frustrated with judicial discretion because the biased attitudes of some judges shaped their decision-making, leading to disproportionate sentences for minorities. Meanwhile, in light of increased drug use in cities, there was a push to create laws that would dissuade drug use. Mandatory-minimum sentencing for drug crimes was introduced in the mid-80s with the hopes that it would both end racially biased sentencing and reduce drug use. This backfired. Mandatory sentencing, with its harsh minimum terms, ushered in the era of mass incarceration, a burden that falls disproportionately on communities of color.

Given the problematic use of force by police—and the failure of current policy to stop it—many advocates want to use the tools of surveillance to create oversight that will sharply limit police discretion. The logic is that if police officers are being watched, they will not stray from protocol.

There’s a long history of workplace surveillance being used to increase employee compliance, but this doesn’t always produce desirable outcomes. Part of what’s at stake depends on the context in which the surveilled employees are operating and the various incentives that shape their practices. Law-enforcement officers are unevenly distributed in cities, with many more placed in low-income communities. Discretion doesn’t just let officers choose when to use force; it allows them to choose when to make, or not to make, an arrest. The idea is that automating their decision-making processes by making them cognitively aware of being under surveillance will improve their behavior and help remove the human bias factor to reduce violence.

Just as intersection-speeding cameras send everyone who speeds a citation, there is the potential that police will use ticketing and arrest far more frequently if they feel as though they might be second-guessed for not making arrests. And because of where they are stationed, this means that low-income communities may see a rise in arrests as well as surveillance, particularly if policing practices are increasingly focused on data-driven outcomes. Policing actions that improve community safety outside of quantifiable metrics, like arrest quotas, are harder to evaluate (and you can’t just buy a new tool to try and achieve them). Perhaps that’s the conversation we should be having.

Being recorded entails an implicit threat that the footage may be wielded against you, especially in the context of a police-made recording. This fear is especially resonant for undocumented immigrants or those with outstanding warrants. Already, many don’t seek help because they fear being identified. The issue isn’t that people don’t trust cameras; it’s that they don’t trust the police, and, by extension, the ways police may use a tool that gives them more power.

New surveillance and data-collection practices in law enforcement have already created frames for broadening police powers, and in particular, for increasing the over-policing of communities of color. These communities are already being affected by predictive policing with heat maps, police surveillance of social media, police collection of biometric information like DNA from people who are merely stopped by police (not arrested), and the proliferation of CCTV cameras on every corner of “high crime” neighborhoods. These are all tools and practices that have been legitimized and popularized through widespread use. They’re also supported by Supreme Court decisions that privilege the increase of surveillance and the power of the institutions that wield them at the expense of the communities whom police encounter most.

As David Brooks argued in The New York Times, body cams will erode trust within society. Although he felt as though this would be a fair trade-off, he noted that “society will be more open and transparent, but less humane and trusting.” The human consequences he articulates—the costs of surveillance on police and families and the destruction of the social fabric through technological mediation—are too often ignored in the emergent debates around police-worn body cams. Regardless of how adversarial the public’s relationship with police already is, a reduction in discretion can backfire and magnify existing divisions.​

* * *

Surveillance has an economic and social price. Advocates hope that the psychological cost of being watched will dissuade officers from abuse, but members of these communities will face these costs too.

Police-worn body cams do not face the police. They face members of the community—everyday people doing everyday things. The goal may be to capture criminal activities by civilians and by police, but to get there, these cameras will film people walking down the street minding their business. What happens when body cams get additional features, like facial-recognition technologies? Advocates imagine that these tools will surveil the police, but unlike most footage that has rallied the public, body-cam footage does not make visible the facial expressions or movements of the police officer’s body because it doesn’t face the officer. Instead, it casts a spotlight on the other people among whom police operate. Worse, it focuses attention on the limited angles that can be seen from that vantage point. As any videographer knows, camera angles matter. You can tell radically different stories depending on the angle. For instance, due to the average heights of male officers and the women they encounter, a lapel camera can capture a steady stream of cleavage when an officer is facing her.

Those who will be surveilled by body-worn cameras are already the most marginalized members of society, and they already experience a disproportionate amount of surveillance from other law-enforcement cameras. This is particularly true for people who lack private residences to retreat to, either because they are homeless or are resident in public housing, where the police have greater freedom to enter over a resident’s objection. Adding more surveillance to an already hyper-surveilled community doesn’t benefit that community. As the renowned urban theorist Jane Jacobs highlighted, safety in a community does not come through bureaucratic processes; it comes when members of the community are collectively looking out for each other through a dynamic that she calls “eyes on the street.” These eyes cannot be replaced by cameras wielded by authorities who are perceived to be oppressive.

There’s a reason that Taser’s stock went through the roof after Michael Brown’s death in Ferguson: Providing these technologies is very profitable to companies and very costly to cities. Even those who accept the social costs of surveillance must consider the fiscal costs. The roll-out cost of body-worn cameras isn’t just the price of the equipment. There’s also the cost of storing and analyzing the footage. There’s the cost of securing the footage, and the personnel cost of making it available upon request. If we know that we can meaningfully increase accountability, spending this money is a no brainer. But we should also be wary of the economic interests that shape so much of our criminal-justice system, especially if these cameras turn out to be ineffective. Too often, our government enters into long-term contracts with vendors before testing out the efficacy of a new technology and then we pay the price both economically and socially. Why aren’t we looking to fund community advocates in lieu of new technologies? Why aren’t we more focused on what is happening with vendors?

In 1986, Melvin Kranzberg argued that “technology is neither good nor bad; nor is it neutral.” He wrote “technology’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.” We turn to technology because it’s new and fancy, because we hope that it can do things that people cannot do. But technology is a tool. And it is primarily used to reinforce existing structures of power, even when we hope that these same tools can be used to challenge the status quo.

Public consensus is clear: Policing practices must be challenged and policing abuses must end. The American people are now immersed in data showing the costs of policing on our society, particularly for communities of color. We also know that, if history serves, the visibility of these endemic issues won’t continue forever. As a result, there’s a need to act and an opportunity to push for change presented by the public’s attention to the deadly use of police force against black men and the #blacklivesmatter movement.

As researchers, we recognize that the urgent push for police-worn body cams stems from this important moment and that pushing for hesitancy, assessment, and a slow roll-out isn’t publicly desirable. Yet after analyzing the dynamics of surveillance and power over many years, we can’t in good conscience celebrate the desire to go full-speed ahead. While we are fully in favor of citizen video as a tool for raising awareness, generating action, and galvanizing communities to push for accountability, we don’t believe that police-worn body cams will achieve anywhere near the same outcomes. Instead, we worry that they will be expensive distractions that are used to do more harm than good.