Q&A with five experts on whether it is time to regulate facial recognition technology and its usage by police, and what needs to be done to overcome racial bias
Facial recognition is everywhere — airports, police stations, and built into the largest cloud platforms in the world …
Context & Ripple Effects
This Q&A lands two months after Kairos CEO Brian Brackeen argued that facial recognition, given its abuse potential and documented bias against people of color, should not be used by police at all an unusually blunt stance from inside the industry. The five experts are being asked to settle what that piece left open: whether the answer is regulation or refusal, and whether bias is fixable.
The debate it frames has only sharpened since — US law enforcement's use of the tech has since been mapped in detail with critics pressing for transparency and government rules, Microsoft and Amazon have flipped from selling the tech to lobbying for federal regulation of it, and a Washington Post investigation found 15 police departments across 12 states arresting suspects identified by face matches alone.
First-order effects
- Police departments using facial recognition face a direct credibility challenge: the experts' bias findings mean every match they act on can be attacked as racially skewed evidence, not just imperfect software.
- Vendors are forced onto the defensive on two fronts at once — civil-society pressure over bias and, for firms like Kairos, an internal industry split over whether police should be customers at all.
Second-order effects
- Once bias becomes the headline, large cloud vendors find it cheaper to lobby for federal rules than to keep defending deployments city by city — which is exactly the pivot Microsoft and Amazon made when local laws started fragmenting the market their push for federal regulation came alongside a wave of local ordinances.
- State legislatures become the testing ground for compromise bills, as with the proposed Massachusetts law that would permit image matching but ban face surveillance and emotion recognition drawing a line between search and surveillance.
Third-order effects
- Even if government use gets strictly regulated, the same tech's ubiquity in consumer devices keeps the underlying capability unregulated — oversight of police use addresses the most visible deployment, not the technology itself as Vox's follow-up argued.
- The pattern holds through 2025: departments were still arresting suspects on face matches alone years after these debates began, suggesting regulation lags deployment unless rules attach consequences to how matches are used as evidence the Post's seven-year-later investigation.
The trend: Facial recognition governance is moving from expert argument to fragmented state and local rules, while police adoption continues to outrun whatever oversight each jurisdiction manages to pass.