On May 25, 2020, the world learned of the brutal killing of George Floyd by Minneapolis police thanks to cell phone footage filmed by Darnella Frazier of Floyd’s last moments. Police-worn body camera footage of the whole incident, later filed in court, with the transcript made public on July 8, contains every bit of the horror of Frazier’s video. The eighty-two-page transcript shows that Floyd told the police at least twenty-seven times that he couldn’t breathe, to which Derek Chauvin responded: “Then stop talking, stop yelling. It takes a heck of a lot of oxygen to talk.”

The footage was filed by Thomas Lane, one of the officers charged with aiding and abetting murder who aims to have these charges dismissed. It remains to be seen whether this evidence will strengthen his case. But it’s clear that its existence is a big positive for Axon, the company whose camera captured it and whose name is plastered at the foot of each transcript page.

During the peak of June’s Black Lives Matter protests, Axon’s share price had risen to its highest ever level of $101, and today, it still stands 10 percent higher than its price the day before Floyd’s death.

Axon’s offerings currently include body and dashboard cameras, streaming drones, cloud solutions to store the petabytes of footage this hardware creates, as well as AI tools to analyze this evidence and even carry out automated live crowd monitoring. The company is just one of the larger players in the law enforcement software market, estimated to be worth around $10 billion, which also includes tech giants like Amazon and Microsoft.

But none are as evangelical about the promise of digital policing than Axon’s founder and CEO Rick Smith. Smith has promised to the media on numerous occasions that his company would help “make the bullet obsolete . . . it will change the world when we do it.” He even wrote a book on this topic last year called The End of Killing: How Our Newest Technologies Can Solve Humanity’s Oldest Problem. Axon’s website makes similarly lofty claims, estimating that, to date, 225,000 lives have been saved by their products.

Smith has been somewhat more reluctant to acknowledge that the sole technology the company was set up to sell in 1993 can be every bit as deadly as the gun. Trading as Taser International until the company rebranded in 2017, Axon has provided hundreds of thousands of police officers across the globe with a tool estimated to have killed at least five hundred people in the United States and Canada over a ten-year period.

Despite this, much of the current energy directed at severing links between police and tech companies has largely bypassed Axon and focused primarily on ensuring that giants like Google, Amazon, and Microsoft no longer sell products or enter into contracts with police forces. These companies’ workers emphasize the hypocrisy inherent in issuing statements saying they support Black Lives Matter while still providing tools for racist surveillance.

But it’s perhaps unsurprising that companies like Amazon adopt the same moral principles they hold in every market they attempt to colonize — until public relations concerns make it no longer tenable for them to do so, at least temporarily, with their announced one-year moratorium on selling facial recognition technology to the police.

The case of Axon, however, shows how some tech companies can become much more deeply embedded in police forces, remaining immune to criticism and profiting from cases of police brutality — all the while maintaining an outwardly progressive social mission that keeps the promise of moderate police reform alive while skirting calls like “defund the police.”

By the early 2000s, Axon was facing bankruptcy and its business was not growing significantly. That changed when it pivoted heavily into providing body cameras in 2008. As with its current rise in profitability, a police shooting almost a decade ago enabled Axon to promote its new product. In November 2009, police officer Brandon Davis shot and killed Eric Berry inside his own home. Footage from the incident, though blurry and unclear, was enough to ensure charges weren’t pressed and allow the local prosecutor to conclude that Davis’s use of force was justified.

Axon publicity materials readily began featuring this case. Some emphasized the speed of his exoneration: “He was back on the job just 72 hours after the shooting!” Others used pictures and quotes from Davis himself about his enthusiasm for the company’s cameras. In lieu of any clear research on the impacts of body-worn cameras at this point, they were sold as a way to make policing safer for police and provide evidence if needed after the fact.

Michael White, coauthor of Cops, Cameras, and Crisis, was approached by the Department of Justice several years later to produce the first systematic review of the evidence around body-worn cameras. “It was an easy report to write because there were only five studies that had been done at that time. The report was published in April of 2014, and I don’t think anyone read it until August 9, 2014, when Michael Brown was killed,” says White.

By December 2014, President Barack Obama was planning to set aside $263 million to fund body-worn cameras and police training, a large amount of which was funneled into Axon, already the camera of choice for police. The continued police killings and cases of police brutality should be evidence enough of their failure to produce reform, but a 2019 meta-analysis also found that body cameras can only ever be as effective as the departments they’re used in allow them to be.

“One study looked at the Phoenix Police Department, and the activation rate was about 30 percent, which means that 70 percent of the time, when there was supposed to be a video, there wasn’t. I’ve also seen some studies that have looked at individual officers in the department, with activation rates as low as zero or 1 percent,” says White of research produced since his 2014 report.

Despite this, millions of dollars are still being paid out by police forces to companies like Axon, who are aiming to secure steady incomes and shield themselves from calls to defund police departments by operating as a “software as a service” (SaaS) company, locking forces into five-year contracts for their Officer Safety Plan, providing hardware and cloud storage facilities for $109–$199 per officer per month. The plan promises to help forces “do more with less,” a familiar offering from tech companies to public services facing potential budget cuts.

This often involves outsourcing digital storage, providing algorithmic data analysis, and, disturbingly, leaning on the public to join in on digital surveillance. Citizens can now seamlessly upload footage to Axon’s evidence cloud server, along with racially profiling their neighbors on Nextdoor or providing footage to police officers via their Amazon Ring doorbells.

Calls to defund the police should ensure public funds no longer go toward unaccountable tech companies nor expanding digital surveillance, and we should be wary of the risk of these surveillance tools being picked up instead by members of the public. When the city of Camden, New Jersey, disbanded its police force, in its place came hundreds of cameras, ShotSpotter microphones, and a mobile observation tower mostly operated by civilian contractors.

The Open Tech Institute, based in Washington, DC, has worked to introduce legislation and bans on many of these surveillance tools for years. It is currently part of the “Community Oversight of Surveillance” campaign, which is pushing for the DC city council to take up transparency and oversight legislation on surveillance technology. Surveillance planes flying over the BLM protests in DC during June highlight how interlinked ending police brutality, oversurveillance, and abuse of technology are as demands.

“Surveillance is deeply intertwined with policing,” says Lauren Sarkesian, senior policy counsel at Open Tech Institute. “Oversurveillance leads to disproportionate arrests, causing more black and brown people to enter the criminal justice system, as they are more extensively surveilled. So these issues of police surveillance technologies including facial recognition, body cameras, drone surveillance, CCTV cameras, IMSI catchers, and more should be considered and overhauled.”

Inevitably, communities will be safer when their residents are not targets for racial profiling, and when their streets are not data points for predictive policing tools. Violence at the hands of the criminal justice system can be reduced when there are fewer police on the streets, but also fewer drones and surveillance cameras above them. In the last five years, Minneapolis police ran training on everything from implicit bias to mindfulness, and their complete failure highlights the empty promise of even the mildest liberal reforms. We can’t lose sight of the tech companies invested in keeping that empty promise alive.