Artificial Intelligence accused of targeting people of color

Our Take

Credit to InvestigateTV for digging into wrongful arrests tied to facial recognition technology and the research showing these systems misidentify people of color at higher rates. This isn't a new story, but it's one that keeps repeating because nobody with the power to stop it wants to slow down the rollout. Facial recognition and automated license plate readers come from the same playbook: sell police departments on a shiny AI tool, promise it'll only be used for 'serious crimes,' then watch the technology creep into everyday policing with almost no independent audit of how often it gets things wrong.

Flock Safety and its ALPR network don't identify faces, but they operate on the same flawed premise — that an algorithm's output is trustworthy enough to justify a stop, a search, or an arrest. When a plate reader misreads a character or flags a rental car as 'associated' with a suspect vehicle, the person behind the wheel pays the price, just like the people InvestigateTV highlights who were picked up because a computer was confident and wrong. Bias baked into training data doesn't stay contained to one technology; it's a feature of the entire automated policing ecosystem, and ALPRs are quietly becoming one of its biggest pillars.

The lesson here isn't just about facial recognition — it's about what happens when we let machine judgment stand in for human accountability. Check our camera map to see how much of this infrastructure is already watching your community, and visit our take-action page if you want to push back before the next wrongful stop happens near you.