The footage from the evening of May 25th in Minneapolis was filmed by 17-year-old Darnella Frazier on her phone. Sixteen times over five minutes, George Floyd says he can’t breathe. A group of onlookers, including Frazier, hurl anguished pleas at the officers, including Derek Chauvin, a two-decade veteran of the force, who keeps his knee on Floyd’s neck for nearly nine minutes. “I’m about to die,” Floyd says at one point. Chauvin tells him to relax.
Frazier’s video has sparked worldwide protests over a long history of brutal, biased policing, and underscored the profound power of video. Without the footage, Floyd’s death might have been what the Minneapolis police initially described in a statement as simply a “medical incident during a police interaction.” The press release made no mention of use of force.
Fortunately, Frazier wasn’t the only one recording: the officers were also filming the entire encounter on their body cameras, the result of a previous round of reforms aimed at reducing force and enhancing transparency. And yet, the public still hasn’t seen those videos: Like many states, Minnesota gives police wide discretion about when and how to release the footage, if at all.
It’s a pattern repeated at police departments across the country, and it adds to a growing chorus of questions about the actual impact of police video.
“We spent a king’s ransom on body cameras in this country, for accountability,” says Barry Friedman, a professor at New York University School of Law, where he is director of the Policing Project. “But without a policy in place to release footage to the public, they simply disappoint the community and create further tension.”
We spent a king’s ransom on body cameras in this country, for accountability.”
Police in almost all of the country’s large police departments now wear cameras, as part of one of the swiftest, costliest adoptions of new police technology since the walkie-talkie. The upgrade began in earnest after the fatal shooting of 18-year-old Michael Brown in 2014, which was not caught on video. In 2015, the Justice Department and states began giving police departments millions of dollars to buy cameras as part of a flurry of reforms meant to improve police interactions and provide more accountability: revised use-of-force policies, anti-bias and deescalation training, federal oversight of troubled departments.
Even as some Obama-era reform efforts were scrapped under a new administration, body cameras have remained a stalwart plank in local and federal reform efforts. Police reform bills unveiled by Senate Democrats and Republicans would plow more federal funds into the devices. Despite calls to review big police budgets and overhaul policing, citizens are still clamoring for them. Shareholders have taken stock too: Axon, the company that sells body cameras and Tasers to Minneapolis and many major departments, has seen its share price surge by as much as 25% since Floyd’s death.
This hasn’t been cheap. Police departments, through state and federal grants, have spent upwards of $100 million on cameras. (The Dept. of Justice alone has given $73 million to more than 400 agencies from 2015 to 2019.) That’s meant a gold rush for startups and companies like Axon, but also Amazon and Microsoft, who have built lucrative subscription-based businesses around storing mountains of body camera video. At some smaller departments, The Washington Post reported last year, the rising and recurring costs have led them to scrap their camera plans altogether.
However, when it comes to reducing police use-of-force or building community trust, the video experiment has not had the effect many reformers had hoped for. Like citizen-shot video, police video is still hindered by a larger operating system: police training, practices, and cultures that promote aggressiveness and exacerbate existing racial biases, as well as old laws and police unions that foil transparency and accountability. Body camera video, meanwhile, can also be combined with live-streaming and real-time face recognition, technologies that can harm privacy and civil rights and reinforce racial disparities.
With existing policies and emerging technologies, body cameras “quickly have become yet one more tool for surveillance and collecting evidence for prosecution,” said a panel of veteran criminal justice scholars in a report last week.
Ethan Zuckerman, former director of the Center for Civic Media at the Massachusetts Institute of Technology, was one of many who hoped that cameras could stem disproportionate use of force against black men. But, he wrote last week, that was “a techno-utopian fantasy . . . a hope that police violence could be an information problem like Uber rides or Amazon recommendations, solvable by increasing the flows of data.” But, he wrote, it’s “clear that information can work only when it’s harnessed to power.”
As the George Floyd protests have shown, video matters. But how much it matters, and for what, depends on who is filming it and who ultimately gets to see it. Understanding the actual impact of body cams—and how they haven’t had the effect many had hoped—can help guide ideas about future reforms.
Improving police interactions: The jury is still out
As with police use of force, there is no national database on how police use body cameras. To better understand their impact, researchers at George Mason University’s Center for Evidence-Based Crime Policy instead looked at 70 academic studies on U.S. body cameras programs between 2015 and 2018. In a paper published last year, they reported that while police and the public generally like cameras, the devices have had few significant effects on most measures of officer and citizen behavior or citizens’ views of police.
[I]nformation can work only when it’s harnessed to power.”
“Increasingly, officers value BWCs”—body-worn cameras—”as a tool for their protection (against false or exaggerated accusations of wrongdoing), for evidence collection (which may be bolstered by prosecutors’ support for BWCs), and for accurate reporting,” write the researchers, led by criminology professor Cynthia Lum. “It may be fair to say, however, that BWCs have not produced dramatic changes in police behavior, for better or worse.” Early findings had indicated that cameras reduce the use of force by officers, but more recent findings were mixed, “perhaps in part as a result of variation in agency policies regarding how the devices should be used.”
Carley Partridge, a spokesperson for Axon, the Scottsdale, Arizona-based body camera giant, said that cameras “cannot take the place of good policies, training, and community policing efforts.” Despite the GMU researchers’ findings, she contends that “the majority of studies” show body cameras have had “a positive effect” in reducing use of force and increasing transparency. The notion inspired the company’s slogan—”Protect Life, Protect Truth.”
Cameras may be more effective in certain places, or in certain units within police departments. In a study published last month, researchers found that cameras had no impact on use of force among patrol officers in Tempe, Arizona. But among officers in specialized units—which, the researchers noted, use substantially more force than patrol officers—cameras “were associated with a significant decline in force.” Special units are not always included in body camera roll-outs: While patrol officers in Louisville, Kentucky, are required to record all public encounters on their body cameras, the plainclothes narcotics officers who fatally shot Breonna Taylor in her bedroom last month were not wearing cameras.
Body cameras have also helped police officers gather evidence, and in some cases, may help reduce complaints by citizens. In Henry County, Georgia, where the body cameras are modified Android phones sold by a startup called Visual Labs, police spokesperson Randall McGlamery told me that body-worn video has proved useful, “in terms of complaints and use-of-force reports, and getting to the truth and understanding what happened quicker, and putting to bed any false allegations from citizens.”
A number of studies have also reported a decline in citizen complaints for officers wearing cameras. Still, it’s not clear if that trend reflects changes in police behavior or in citizens’ reporting behavior, or both. It’s not even clear if complaints themselves are a useful way to measure officer behavior or officer-citizen interactions. More research is needed to understand how cameras impact those interactions, says the GMU study, “such as systematic social observations, ethnographies, and even analysis of BWC footage itself.”
In some cases, cameras have also caught nonviolent misbehavior by officers, like planting evidence or even trying to manipulate the record itself. In two separate incidents in 2016, Baltimore detectives were caught on their body cameras planting drugs on or near suspects, before attempting to use the cameras to substantiate what they’d found. The release of the videos inflamed longtime public frustrations, and led the state’s attorney to throw out dozens of criminal cases.
When police control the video
As a transparency tool, body cameras fail when officers manipulate what’s on video—or if they simply don’t record, or lose their cameras, or the video.
“Some of the most important videos of officer misconduct have come from citizen recordings, while police-worn body cameras often were not activated or only caught part of an event—or, indeed, were used to create a narrative at odds with the facts,” says Rachel Levinson-Waldman, senior counsel to the Liberty and National Security Program at the NYU Brennan Center.
Even when they’re used as intended, the devices are limited by police policy. When are officers required to turn them on, and what penalties will they face if they don’t? Are police allowed to review video before they write incident reports, which can allow officers to more easily tailor false accounts to the footage? How do videos get released to the public?
The answers to these questions vary depending upon where you are, the result of a patchwork of state and local laws that are crafted with significant influence from powerful police unions. Some states consider body camera videos to be public records, and departments with histories of controversial encounters, like Chicago and Los Angeles, mandate that videos of police shootings be released within days. If they do release video publicly, departments will first blur faces to protect the privacy of citizens who appear in the videos, including alleged victims, perpetrators, bystanders, and protesters.
But many localities can keep video of violent, controversial encounters hidden from view, either by citing privacy laws or classifying the videos as evidence in ongoing investigations. As a result, prosecutors and reporters must often resort to lawsuits to see critical body camera footage.
Since New York City began its body camera program in 2017, videos have only been released at the discretion of the commissioner. Last year, the New York Lawyers for the Public Interest sued to obtain unredacted footage from the first fatal incident captured on NYPD cameras, the shooting of a 31-year-old exchange student from Jamaica named Miguel Richards. Richards was experiencing a mental health crisis in his Bronx apartment when officers arrived. They asked him 41 times to drop a knife and later, what appeared to be a gun, before killing him in a hail of gunfire.
The NYPD had released an edited version of the footage, but refused to release more. Disclosing more footage, said Patrick J. Lynch, the president of the city’s largest police union, “jeopardizes police officers’ due process rights and confidentiality protections under state law.” A judge ruled otherwise, and ordered the NYPD to release the video.
In theory, videos like these illustrate the value of body cams. They make possible a clear-eyed review of bad encounters, and help show why, for instance, the public should control the footage.struggle to pry the videos from the department shows how limited body cameras can be, and why some reformers think
In recent weeks, lawmakers have begun chipping away at laws and policies that make footage so hard to get. On Monday, New York Governor Andrew Cuomo repealed an old rule that shielded officers’ discipline records from the public. On Tuesday, New York City Mayor Bill de Blasio announced a sudden shift in NYPD policy: After any incident where officers use force and someone is badly injured or killed, body camera footage will now be released within 30 days.
“We’ve focused on the power of transparency with body-worn cameras,” de Blasio said at a briefing. “But it only works if people see accountability, see results from the presence of those cameras.”
Still, it wasn’t clear how the policy would impact how the NYPD redacts footage before releasing it. According to a City Hall spokesperson, the department will retain the right to edit the videos for “privacy.”
Video isn’t the whole story
Videos can be a powerful witness to the facts. If a bystander hadn’t been recording, the police record would have shown that the 75-year-old man shoved to the pavement by a Buffalo police officer this month “tripped and fell.” It’s a now familiar pattern: Video at times sharply contradicts statements by police or officials.
But video isn’t always a reliable witness. Even unedited body video may offer only what the policy watchdog Upturn calls an “illusion of accuracy.” For instance, the framing of a body camera lens can exaggerate the danger that an officer faced during a violent encounter. This effect of “deceptive intensity” can play a significant role in arguments of qualified immunity, the controversial legal doctrine that shields officers from liability over the use of deadly force, provided they have an “objectively reasonable” fear that their safety is in danger. Criminal charges against police, and especially convictions, remain exceedingly rare.
The immunity defense—and the body camera footage—is likely to play a role in the trial of the officers charged with aiding and abetting the killing of George Floyd. In an interview on CNN last week, an attorney for Thomas Lane, one of the Minneapolis officers who helped pin Floyd to the ground, said he had seen some of the body cam video, and suggested that it was exculpatory.
The video captured the beginning of the encounter, and showed the critical perspective of the officers. Floyd wasn’t violently “resisting,” the lawyer said, but “it was not a kind of nonresistance that an individual should do when a police officer is arresting him.” It may not change the basic brutal facts, but in court, that kind of video evidence can be advantageous to officers.
What gets recorded and what doesn’t matters too. New York City’s Civilian Complaint Review Board, which uses body camera videos to investigate allegations of police misconduct, said that it often heard NYPD officers using phrases like “I went Hollywood,” “Green,” “We’re live,” and “I’m hot,” as well as nonverbal cues, to warn fellow officers that their body-worn cameras were recording. This phenomenon, the review board wrote in a February report, “undermines the purpose of the BWC program, which is meant to, ‘provide a contemporaneous, objective record of stops and frisks, allowing for the review of officer conduct.'”
New surveillance powers
Rather than fueling transparency, many civil rights advocates have long worried that police could use cameras to supercharge surveillance, for instance, by scanning the faces in crowds for criminals. In 2015, the U.S. Justice Department warned police that the technology could harm Fourth Amendment protections and lead to biased and dangerous misidentifications.
To many in the police industry, however, the technology is seen as inevitable, part of a high-tech vision of policing evocative of Robocop and Minority Report. Since 2009, Axon has speculated about using its body cameras to live-stream video and identify people in real time; and in 2018, it acquired a patent along those lines. The company placed a temporary moratorium on the technology after its ethics board raised concerns, but amid a glut of face recognition software and giant face databases, its vision is steadily coming into focus nonetheless.
No federal laws govern face recognition, but some cities have demanded more transparency over police technologies. Some, including San Francisco and Cambridge, Massachusetts, have even banned the police use of face software. The police reform law introduced this week by House Democrats would also place federal restrictions on the technology. Last week, IBM, Amazon, and Microsoft also said that they would stop or limit sales of the technology to police.
In theory, video analysis software could also be a tool for accountability. AI can help researchers more easily glean data from piles of body camera footage, or provide police departments with an “early warning system” to help detect signs of bad officer behavior. Axon says a future tool will be able to scan footage to flag certain events, like an officer using profanity, raising their voice, or activating their Taser. Still, the effectiveness of those systems will depend on the department using them. For instance, by the time Officer Chauvin killed George Floyd, the Minneapolis Police Department knew that the officer had already racked up a remarkable 18 complaints and three shootings on his official record.
Without stronger rules, critics say cameras and other high-tech approaches to reducing police violence are futile, or worse: They can enhance surveillance and distract from other effective reforms.
In a study published last year, Rutgers sociology professor Michael Sierra-Arévalo described how new policing tools like Tasers can sometimes contribute to the very problem the tools are intended to address. Stemming brutality would require addressing “persistent features of police training and culture that socialize officers into an ‘us versus them’ orientation that frames the public as potential threats instead of fellow citizens and allies.” In the meantime, the videos will keep offering gruesome reminders of the reforms that remain unfinished.