Axon Ethics Board Recommends Against Facial Recognition in Body Cameras, Company Agrees to Pause Development

On June 27, 2019, Axon announced it would not commercialize facial recognition technology for body cameras following recommendations from its AI and Policing Technology Ethics Board. The board’s first major report concluded that “facial recognition simply isn’t good enough right now for it to be used ethically” and warned that deployment of the technology in law enforcement contexts posed unacceptable risks of racial bias, privacy violations, and erosion of civil liberties. Axon agreed to pause development of face-matching capabilities for its body camera systems, marking a rare instance of a major surveillance technology company accepting limitations on profitable AI applications due to ethical concerns.

Ethics Board’s Six Key Findings

The AI and Policing Technology Ethics Board, chaired by NYU Law Professor Barry Friedman, issued six critical recommendations regarding facial recognition in police body cameras. The board emphasized that companies should report specific false positive and false negative rates rather than misleading general accuracy metrics, warned that overly customizable facial recognition systems could be manipulated to amplify bias, and insisted that any deployment require meaningful consent from affected communities with strong evidence of genuine public safety benefits. The board distinguished between face detection (identifying facial features) and face matching (comparing faces against databases), noting that the latter posed far greater risks to privacy and accuracy. The board further emphasized that facial recognition systems operate within political and social contexts where they could cause disproportionate harm to overpoliced communities, making technical solutions insufficient without addressing systemic justice concerns.

Company Response and Significance

In response to the ethics board’s findings, Axon committed to not proceeding with development of face-matching products, including adding these capabilities to body-worn cameras. The decision was significant given Axon’s dominant market position: an estimated 70-80% of major US police departments used Axon body cameras, meaning the company’s choice effectively prevented widespread deployment of facial recognition in police body cameras nationwide. The pause was framed as temporary (“for now”), with Axon stating it would reconsider if the technology improved sufficiently to meet the ethics board’s standards. Civil liberties advocates praised the decision as an important precedent demonstrating that ethics boards could meaningfully constrain corporate behavior in the surveillance technology sector, though some noted the “for now” language suggested commercial considerations might eventually override ethical concerns if the technology became sufficiently profitable.

Technical and Bias Concerns

The ethics board highlighted that facial recognition algorithms demonstrated documented patterns of racial and gender bias, with significantly higher error rates for women and people of color—precisely the populations most likely to be subject to intensive police surveillance. The board noted that even “accurate” facial recognition systems could enable mass surveillance capabilities incompatible with democratic norms, allowing police to identify and track individuals in real-time as they moved through public spaces. The integration with body cameras was particularly concerning because it would enable retroactive identification of everyone captured in police footage, creating a comprehensive database of public movements without consent or probable cause. Technical experts on the board emphasized that improving algorithmic accuracy would not address the fundamental civil liberties risks posed by pervasive facial surveillance.

Contrast with Later Ethics Board Collapse

The 2019 facial recognition decision was later cited as evidence that Axon’s ethics board could successfully influence company policy when the company chose to listen. However, this successful collaboration stood in stark contrast to the board’s June 2022 collapse, when nine of thirteen members resigned after Axon ignored their recommendations and announced plans to develop Taser-equipped drones without consulting the board. The facial recognition episode demonstrated both the potential and the limitations of corporate ethics boards: they could be effective when companies voluntarily accepted restrictions, but lacked enforcement power if corporate leadership chose to override their recommendations for commercial or strategic reasons.