Why Body Camera Programs Fail

Amid a massive tech upgrade meant to improve accountability, body cameras are being limited by police policies and privacy concerns.

Why Body Camera Programs Fail
[Photo: Flickr user West Midlands Police]

Read more stories about body-worn cameras and the Future of Policing.



In the minutes before a police officer in Baton Rouge, Louisiana, killed 37-year-old Alton Sterling as he lay restrained in a convenience store parking lot, both the officer and his partner’s body cameras “came dislodged” from their uniforms.

A month later, in August, a Chicago officer failed to turn on his body camera before he fatally shot unarmed 18-year-old Paul O’Neal in the back.


And a few weeks ago in Charlotte, North Carolina, the plainclothes officer who fatally shot 43-year-old Keith Lamont Scott wasn’t wearing a body camera. The other officer involved in the incident only activated his body camera after the shooting, a violation of city policy.

A body camera video shows the aftermath of the shooting of Keith Scott. The first 30 seconds are silent: The camera, a Taser Axon Flex, captures footage in the 30 seconds leading up to the moment the officer presses record, but this buffer footage contains no sound.

The failures–however accidental or convenient they may be–underscore the limits of a technology that is rapidly becoming a staple of the modern police officer’s tool kit, amid an effort to build more trust between police and the communities they serve.

As she announced $20 million in federal body camera grants last month, Attorney General Loretta Lynch said the money “will help more than 100 law enforcement agencies promote transparency and ensure accountability, clearing the way for the closer cooperation between residents and officers that is so vital to public safety.” Arizona-based police technology manufacturer Taser, whose Axon division now claims to sell the majority of the country’s police body cameras, puts the value proposition more simply in one of its slogans: “Protect Life. Protect Truth.”


But the technology isn’t foolproof. Cameras fall off, officers fail to record, and the video itself can be kept largely out of public view. When video collected for oversight purposes isn’t shared publicly, critics worry citizens might become more suspicious about police misconduct, amplifying mistrust amid an effort to fight it.

“One of the main selling points for body-worn cameras is their promise to bring transparency and accountability to police community interactions,” says Harlan Yu, a founder of Upturn, a Washington, D.C.-based firm that provides analysis to policymakers, and the author of a recent national “scorecard” on body-worn camera policies. But the cameras, he says, “don’t automatically provide accountability.”

Yu points at recent high-profile failures by police to record deadly encounters. “And these are just the incidents that we know about because someone was shot and killed.”


Civil rights advocates have raised concerns about hundreds of known instances in which cameras have malfunctioned, or in which officers simply haven’t activated them during violent encounters. They have also raised questions about the way police manage and share the mountains of video they collect. Police, meanwhile, fret about the privacy intrusions of body-worn cameras–for victims, minors, informants, and witnesses, as well as for themselves–and the added costs and workload required to store torrents of video.

“Some chiefs are really concerned about that and the work that is associated with all of that,” says Seattle’s police chief, Kathleen O’Toole, who arrived in 2014 amid a series of reforms at the police department. “But you know, I think on balance transparency trumps all of the other inconveniences.”

Still, the inconveniences are significant. Seattle, which sought to release much of its body camera video to the public during a trial in 2015, intends to begin deploying cameras to its officers by the end of the year. But the rollout has been delayed by disagreements over the correct policies for cameras, an issue that has riven legislatures across the country. Transparency activists want the cameras always recording; police want to be able to turn them off at their discretion. Some local laws require quick deletion of police video, while others may keep it indefinitely. Some punish officers for failing to record, while others do not. Only 2 of the country’s 20 largest police departments make body camera footage public by default.


To O’Toole, the challenge is finding the right policies to meet the technology. “I’m not as concerned about the technology as I am about the policy, and we just want to get the policies right.”

To build trust with the community, cameras must be implemented carefully alongside other efforts like community policing, de-escalation tactics and training, and public dialogue, says O’Toole. Cameras will be a “valuable tool” for accountability and training, she says. “But I think we have to be realistic: Body cameras are not going to be a panacea.”

The major points of failure revolve around three central questions: how to reduce human or technical errors, when should officers record, and who gets to see the footage. Take the cameras from the shooting of Alton Sterling in Baton Rouge. Though they fell off, according to a police spokesperson, they were still recording video and audio when Sterling was killed. That data has been turned over to the Dept. of Justice, but it remains out of view of the public: A Louisiana state law enacted in August—one of a number passed in recent months around the country—keeps most body camera footage out of the public record automatically.


Technical And Human Glitches

Since the shooting of unarmed, 18-year-old Michael Brown in 2014—an incident that was not captured on camera—activists and city governments have stridently fought for more police oversight, and for what some call the fastest technology upgrade in policing history. “There is a reason the CMPD equips its officers with body cameras,” Justin Bamberg, an attorney for Keith Scott’s family, told a reporter after his death. “Because body cameras provide visual evidence so that when tragic things do happen we don’t have to question exactly what happened.”

Despite reluctance by some police unions and civil rights groups, body cameras have won over most of the chiefs of the nation’s largest cities. While only a handful of the country’s 18,000 police departments wore them before 2012, between 40% and 50% of cops in the U.S. now use body cameras, according to Michael White, a criminal justice professor at Arizona State University who studies the technology. Besides Charlotte, large cities like Chicago (2,000 body cameras so far) and Los Angeles (1,160) plan to outfit their entire forces.

Some police have come to accept these videos—many of them boring, some of them terribly heart-pounding—as a development as inevitable as the camera on civilians’ cell phones, and as a necessary step toward building trust with their local communities. Police officers have also begun to depend on cameras to refute false allegations of abuse, while prosecutors have cited video as a linchpin in a growing number of police misconduct and criminal cases.


But the technology and the policies that surround it are only as effective as the people operating them. And no policy, process, or technology can completely prevent technical problems—or human malfeasance.

In New Mexico, the Albuquerque police department equipped all of its officers with body cameras by early 2011, earning plaudits for transparency. Like most body cameras, the department’s Taser units were designed to record for the entirety of an officer’s shift. Afterward, videos can be set to automatically transfer to a police computer or stored on a cloud server. But for nearly three years, officers could delete whichever recordings they wanted at the end of their shift, before anyone else saw them.

Officials had enabled an “offline mode” feature in Taser’s video management software,, that allows officers to upload footage to a local computer rather than to a cloud server, not unlike the way photos or videos get transferred off a consumer digital camera. The policy was not publicly known until last year, but it reportedly drew interest from officials at police departments across the country. This feature wasn’t unique to Taser’s cameras; in fact, most body cameras on the market, especially earlier versions, only offer an “offline” mode.


“One of the advantages of the way we built our system is to make sure we have really clear audit logs,” says Taser CEO Rick Smith. But those audit logs only exist if the camera and its software is set to “online” mode.

Taser’s software can offer the option to upload videos to a local computer, where they can be deleted, rather than automatically sent to a cloud server.

The option to delete came to light after the police department dismissed officer Jeremy Dear. In April 2014, he fatally shot 19-year-old Mary Hawkes in Southeast Albuquerque, who he testified was armed. But there was no video of the incident captured by his lapel-worn Taser camera. The department and Dear indicated it “malfunctioned,” but a test by Taser determined that the camera was in working order and was either unplugged or not recording when he shot Hawkes. Officer Dear said he pressed the large button on the battery pack to start recording, but blamed the camera’s design.

“This issue with cords coming unplugged . . . I hear from officers all the time the issue with the cabling is a defect,” Dear’s lawyer told the Albuquerque Journal. “I think there’s a design flaw.”


In 2014, the city of Albuquerque reached a settlement with the Justice Department over a pattern of excessive policing. “We found very few examples of officers being reprimanded for failing to record force incidents,” the Justice Dept. report states. Frances Crockett Carpenter, a civil rights lawyer in Albuquerque, points out that only a few of the nearly 40 police shootings in Albuquerque since officers began wearing cameras have been recorded.

“The citizens of Albuquerque, these Taser cameras were bought on their backs, taxpayer money: millions of dollars to get these cameras,” Carpenter told local TV station KRQE last December. “It was gonna help, and we were gonna get those costs back, right? That money was gonna come back into the city because we weren’t gonna have any more lawsuits. And so somehow the citizens were gonna get some of that money back, or it was gonna flush out. Well, it hasn’t.”

There are dozens of other cases. Last November, two deputies with the Alameda Police Department were caught on surveillance video using their batons to beat a car theft suspect in the middle of a street in San Francisco’s Mission District. Eleven officers in all responded but only one activated his body camera, and did so by accident. Three officers were placed on leave, including two who are charged with assault. But no one was disciplined for failing to turn on their cameras because the department’s policy at the time encouraged, but did not require, their use, according to an agency spokesman. Elsewhere:

  • A study of the New Orleans Police Department found nearly 100 incidents where police used force and were wearing body cameras but did not have them turned on.
  • Last September, two Vermont police officers shot and killed a man while wearing body cameras. Neither officer turned them on before the shooting; both were cleared of all wrongdoing.
  • Just before knocking a woman’s teeth out, two Daytona Beach, Florida, officers switched off their body cameras.
  • A 50-year-old Alabama man was armed with only a spoon—”maybe 10 to 12 inches” in length—when he was fatally shot by a Tuscaloosa police officer, who was wearing a body camera he did not turn on. In doing so, the officer violated police policy, but was not punished.
  • In September, police in Washington, D.C., fatally shot Terrence Sterling, an unarmed 31-year-old black man, after his motorcycle crashed into their car. But contrary to District policy, none of the officers at the scene activated their cameras until after the shooting. The footage the city released captures Sterling’s final moments, but the video begins over a minute after shots were fired. The case is being investigated by the U.S. Attorney’s Office. Now, D.C. officers are required to confirm with dispatchers that they’ve switched on their body cameras when responding to calls or interacting with the public.

In Chicago, where the dashboard camera video of Laquan Macdonald’s death was found to contain no audio, police have been regularly disabling their equipment: One review by a police inspector determined that more than 80% of the department’s dash cameras had non-functioning audio “due to operator error or, in some cases, intentional destruction.” Police have even been seen throwing their microphones onto the roofs of their precinct houses. Similar abuse has been reported in Los Angeles.

According to the Brennan Center for Justice at New York University’s School of Law, most U.S. cities, including Chicago, Dallas, Denver, New Orleans, New York, and San Diego, don’t currently specify penalties when officers fail to record.

“Even when department policies look strong on paper, these policies are only as good as how departments enforce them,” says Yu. “If officers aren’t appropriately disciplined when they fail to record required events—even for routine failures—they’ll form bad habits, and we’ll be certain to see more incidents where cameras aren’t rolling when we need them most.”

Body cam video capturing the aftermath of the shooting of Paul O’Neal. Via Chicago Police Department

Technology may provide some assistance: Taser, VieVu, and other camera makers have begun offering a feature that automatically activates cameras when officers turn on a siren or draw a weapon, much the way that most dashboard cameras are designed to activate when the engine is started or a siren is turned on. In the future, this automatic triggering—and even live-streaming back to base—could happen when an officer enters a crime hot spot, leaves a police vehicle, or makes a call for assistance on the radio. This would avoid difficult situations like the one in Chicago, in which the officer, when giving chase, apparently forgot to turn on his camera.

Encryption used by many body cameras also makes it very difficult for videos to be surreptitiously modified while on the camera or after being transferred to a computer or cloud service. (Some of Microsoft and Amazon servers are federally certified to handle police evidence.) But as the case of Albuquerque’s body cameras demonstrated, it isn’t hard for officers to delete footage, intentionally or not, if videos are uploaded to a local computer rather than to a cloud server.

Last month, a police official in Oakland testified that a quarter of all video clips captured by the department’s body-worn cameras during stops and investigations were accidentally deleted two years ago, when city technicians installed a software upgrade. A spokesperson claimed the deletion had not significantly affected any cases.


In Seattle, in-car dashboard camera videos have frequently dropped frames, at times leaving gaps of more than an hour in video records. Police IT workers and forensics experts have worked on the problem since 2014, according to the department, but the problem persists. And in late August, the department announced that over a two-day period it had lost more than 2,000 dashboard camera videos (due to a computer bug) created with camera systems made by Houston-based COBAN International. Chief O’Toole called it “a technology glitch.” “We want people to know that there isn’t anything sinister here.”

Cameras, whether attached to an officer’s chest or head, can also come off in a struggle, says Smith, the CEO of Taser. The company “rigorously” tests its camera attachments, which tend to rely on strong magnets and clips, but the problem of cameras falling off is serious enough that the company has begun to develop stronger, sewn-in attachments for some departments.

Asked about the possibility of two body cameras falling off during the same encounter, as happened in Baton Rouge, Smith said it was “very rare.” The two cameras that fell off weren’t made by Taser but by Motorola Solutions. In a statement, the company said its “engineers have been on-site throughout the pilot to ensure the body-worn cameras meet the police department’s mission-critical requirements,” and that it had offered the Baton Rouge Police Department “various camera mounting configurations.”

One of Taser’s cameras, the Axon Flex, a lipstick-sized camera mounted on an officer’s temple or lapel using strong clips or magnets, is designed to withstand speeds of up to 100 mph. But that doesn’t account for everything an officer encounters. In a video captured last August by an Axon Flex, an officer in Grand Rapids, Michigan, responding to a noise complaint approaches a group of people on a street corner. The conversation quickly escalates and during an altercation with an older woman, the camera comes loose and lands in a neighbor’s yard. It’s discovered, and the neighbors’ efforts to return the camera are chronicled on video, which the department released to news media: an accidental short film about how body cameras fall off.

When attached and working, body cameras are still limited by their hardware and software, by their lenses and microphones. They tend to only capture what is in front of the officer, rather than their wider surroundings. When officers run, the video image shakes nearly beyond recognition. And when encounters become physical, the video can become indecipherable, frustratingly when it’s most needed. Even if the cameras in Baton Rouge hadn’t fallen off, they may not have captured much valuable visual evidence of the struggle that led up to the shooting, says Smith. “I don’t know if you’re ever going to reliably capture that level of detail in the middle of a fight.”

When To Record

In many cities the debate over body camera policies has slowed their adoption. One sticking point, and a point of potential failure, relates to the question of when officers should record. Civil rights advocates tend to argue that unless police record every encounter, cameras could become a tool for abuse rather than reform. Police say that in order to respect the privacy of citizens, informants, vulnerable victims, and themselves, officers should have discretion about when and what to record.

Shankar Narayan, legislative director of the Washington State ACLU, and a member of a state task force on body camera policy, says body cameras present a new kind of privacy threat. Rather than limited views from police cruiser dashboards, the cameras offer up-close-and-personal HD views of the worst days of people’s lives. Some law enforcement leaders also say that cameras recording everything could be dangerous, and out of step with the nuances of community policing.

Still, he says, giving officers discretion about when to record could defeat the purpose of the cameras. “We should have them be able to turn them off on breaks, but when they’re on duty, the cameras need to be on,” he says.

There have been nearly 40 studies on the use of body cameras, including a dozen randomized controlled trials on the magnitude of their effect on policing. Generally, they have noted a decrease in complaints and assaults against officers who wear cameras, as well as reductions in use-of-force incidents. But, according a recent study by Barak Ariel at Cambridge University, use-of-force actually appeared to increase in cases where officers were given full discretion about when and what to record with their body cameras.

Some departments, including Memphis, allow officers to turn cameras off at their discretion. In Oakland, reprimands for officers who don’t record their encounters with the public are often informal, at least for first-time offenders. Last year, out of 504 use-of-force incidents, 24 officers did not turn on their body cameras. Twenty officers were given reprimands from supervisors and the other four were sent to internal affairs for investigation. Since its camera policy was put into place, the department says it has had a 95% success rate in recording use-of-force incidents.

While New York lacks a statewide camera law, the State Municipal Police Training Council crafted a model policy last fall that states that officers are required to activate cameras “upon engaging in an enforcement-related activity.” But recording wouldn’t be required if turning on a camera placed an officer at a disadvantage, compromised safety, or delayed a response to a citizen in need.

At the NYPD, which recently announced plans to buy 1,000 VieVu body cameras, officers would not record when “interviewing the victim of a sex crime” or when conducting strip searches of suspects.

They would also not record during public demonstrations, in order to prevent against surveillance of political activity. But in theory, not recording during demonstrations could leave numerous contentious incidents unrecorded. (Under an earlier policy in Seattle, for instance, officers could turn on their cameras at demonstrations if they suspected that a crime was in progress.)

Civil rights groups focus on certain other parameters for body cameras, like those enumerated by Craig Levine, an attorney with the Bronx Defenders in New York City: Officers must write an official account of a stop or arrest before reviewing the video, lest they tailor their reports to the footage. Footage must also be provided to the prosecution and defense at the same time, to ensure fairness. And meaningful penalties must be imposed whenever an officer fails to comply with the body camera policy. None of these elements, Levine noted in a letter to the New York Times, are currently included in the draft New York Police Department policy for its body camera program.

In its decision to award grant money to help local departments buy cameras, the Justice Department ensures that police applicants have or are developing a policy regarding their use, which “should at a minimum increase transparency and accessibility, provide appropriate access to information, allow for public posting of policy and procedures, and encourage community interaction and relationship building.” But the Justice Department does not offer specific guidance on how policies should be designed, only specifying that departments must have “a strong plan for BWC implementation and a robust training policy.”

Who Gets To See The Footage?

Even when cameras work and footage exists, a growing number of local laws are putting restrictions on the release of footage. Police departments and state legislators are “making it more difficult for the public to get access to footage of critical use of force incidents or potential complainants to get access to that footage,” says Yu, of Upturn. The effect, he and others argue, threatens to negate the transparency promised by the cameras.

Of the 20 largest police departments in the U.S., just two—Las Vegas and Chicago—have designated body camera video as public by default. Chicago police are required to release footage within 60 days, while Las Vegas typically releases video within 72 hours. And only Las Vegas and Washington have clear procedures in place for individuals who are alleging misconduct to review police footage. Meanwhile, two departments, Los Angeles and San Antonio, don’t release such footage as a matter of practice.

So far, according to a survey by the National Conference of State Legislatures, five states make body camera video part of the public record. The states—Connecticut, Nevada, North Dakota, Oklahoma, and Texas—also provide caveats that allow police to withhold or obscure videos, measures that police and civil rights groups say are essential to ensure privacy. Meanwhile, eight states exclude video from public record requests outright, though some include exceptions for police shootings. Six states have not addressed the issue of public records, and 31 states have yet to finalize laws surrounding body camera footage.

New York Police Department’s current draft policy for body cameras outlines numerous circumstances under which the police could refuse to release footage to the public: if it “interferes with active law enforcement investigations” or “would constitute an unwarranted invasion of personal privacy.” In the case of controversial incidents, the policy states that the police will most likely consult with prosecutors “about releasing the video to the public in order to balance the public’s right to information with the integrity of any criminal investigation or criminal prosecution.” The department’s policy is being crafted with feedback from civil rights lawyers and the public via an online survey, and will be reviewed by a federal court.

In Minnesota, the state will make footage of use-of-force public only if the force “results in substantial bodily harm.” Video may be released in certain cases in which an officer discharges a gun or if the subject of a video requests that it be public. And footage can be redacted or withheld if it is “clearly offensive to common sensibilities.” The rules don’t apply to dash camera footage, which remains open to the public. And the law gives local departments discretion about whether to allow officers involved in a shooting to give a statement before or after reviewing the footage.

Under Louisiana’s new records law, footage from body-worn video is exempt from disclosure if the agency in charge of the record determines that releasing it would “violate an individual’s reasonable expectation of privacy,” or if it is part of an ongoing investigation. Chad Marlow, the national ACLU’s advocacy and policy counsel, expressed concern about letting the police decide whether a video would violate a civilian’s privacy. “That’s a little bit like the fox guarding the henhouse,” he told the Reporters Committee for Freedom of the Press. “It’s an ambiguous standard being determined by a non-neutral arbiter.”

In North Carolina, where the police department gradually released all of its footage of the shooting of Keith Scott, a new law that went into effect October 1 automatically removes such videos from public records. Instead, only people whose voice or image appears in the footage can request to see the video, and prohibits them from copying or sharing the video. It also lets departments deny requests for a variety of reasons, including concerns about the release of sensitive information and the need to protect “active or inactive” investigations or potential future investigations, a reason that is regularly cited around restricting police footage. If denied, a requester can appeal to a court to demand its release.

“Privacy needs to be protected,” says Yu, “but if it requires a huge amount of time and resources or you need a lawyer to get footage in an incident, I think that’s problematic and it raises questions about the extent to which cameras can reach its promise of transparency and accountability.”

Yu says cameras can sow distrust if they aren’t used for transparency. In Chicago and Charlotte, North Carolina, police initially resisted releasing videos of fatal shootings because they were concerned doing so would harm an ongoing investigation and incite public anger. The delays led to widespread suspicions, a public outcry for the video, and violent protests in both cities.

One method for minimizing the privacy risks and costs of releasing video while enhancing transparency is overredaction: A police department publishes all of its videos online, but uses software to automatically blur them to protect people’s privacy. A civilian can then request a section of specific unredacted video. Despite some opposition, the idea became the centerpiece of Seattle’s body camera pilot program last year, after an activist bombarded the police department with bulk requests for video; blurred videos from the small camera trial were uploaded to the department’s YouTube channel, in an experiment that drew the attention of police across the country. The department is determining how it will release body camera footage in the future, but a spokesperson said that it hopes to release video of police-involved shootings within only 24 to 72 hours after such an incident, as it does with dash camera footage.

[Photo: Flickr user Mike Tigas]

Third-Party Video Control

While departments insist they should have control over the video they capture, some civil rights advocates say that footage belongs to the public, and should at least be controlled and distributed by a third party, like the police oversight boards implemented in a handful of cities. Such a system, says Loyda Colon, spokesperson for Communities United for Police Reform, would help to ensure “that there is no tampering and access is available to the public without arbitrary restriction.‎”

But most police say they are opposed to that idea, arguing that body camera video should be controlled like any other police evidence. “Agencies may say they don’t want to release it to the public because it’s counterproductive or incendiary,” says Taser’s Smith. He noted some departments have considered the idea of letting controversial police video be handled by “a committee that represents the people” so that video can be reviewed and released to the public in a timely fashion, without the appearance of a conflict of interest.

Smith says that Taser “provides some guidance” to police customers as they craft body camera policies, and recommends that departments store video on “third party” cloud services. Taser sells storage on Microsoft Azure and Amazon Web Services servers, which are certified to meet FBI security requirements that govern criminal justice data.

“If people don’t trust the police—if they think, ‘well, the police IT department may have gone in and modified the video’—that’s why having that in a third system is a good fit for a law enforcement agency,” he says. Taser, Smith added, “never touches” a department’s data.

Given all of the hurdles that body cameras promise, some critics worry that cameras could be a high-tech distraction from more pressing policing needs. In a memo to local police, the Justice Dept. warns that the deployment of the technology “by itself cannot alter law enforcement–community relations, especially if those relationships have been characterized by long-standing tension and anger. Camera deployment cannot replace community policing.” Some critics have said that body-worn cameras aren’t only a time sink for reformers and a waste of money for communities, but could potentially cause harm, especially in communities where police-community relations are already strained.

“Inserting body-worn cameras in an already corrupt system of policing is harmful. Body-worn cameras mean reinforcing the surveillance apparatus of the state,” New York City organizer and educator Mariame Kaba said in an interview with MTV News. “The resources that are invested—in buying more cameras, paying for systems to retain [footage], etc. . . . takes away from other social goods.” Instead, more citizen footage—and better access to an increasing wealth of surveillance video—could be more valuable for stemming police abuse.

While signing the new body camera bill in North Carolina, Governor Pat McCrory, a Republican, praised the new gear, but also acknowledged its limitations. “When used by itself, technology can mislead and misinform, which causes other issues and problems in our community,” he said, “so what we need to do is walk that fine line.”

Related Video: are police body cameras worth the cost?


About the author

Alex is a contributing editor at Fast Company, the founding editor and editor at large of Motherboard at Vice, and a freelance writer and producer with a focus on the intersections of science, technology, media, politics, and culture.