False Recognition: Racial disparities emerge as new technology incorrectly identifies suspects

Our Take

Credit to InvestigateTV for putting names and faces to what's usually discussed in the abstract: when automated identification tech gets it wrong, real people get arrested, jailed, and traumatized — and the errors fall disproportionately on Black Americans. This isn't a bug in one company's software. It's the predictable result of building law enforcement workflows around algorithmic "matches" that officers treat as gospel instead of a starting point for actual investigation.

Flock Safety's ALPR cameras aren't facial recognition, but the underlying problem is identical: a machine flags something as suspicious, a cop acts on that flag with unearned confidence, and due process becomes an afterthought. We've documented plenty of cases where Flock's own "hot list" alerts and vehicle-description matching send officers swarming on innocent drivers over a algorithm's best guess. Bad matches don't stay evenly distributed — they concentrate in over-policed communities that already bear the brunt of surveillance.

The bigger issue is that neither facial recognition nor ALPR networks were ever put to a real public vote. Departments quietly bolt these systems on, and the burden of proving your innocence falls on whoever the machine points at. Check our camera map to see how much of this infrastructure already surrounds you, and head to take-action if you want to push back before your town becomes the next cautionary tale.