Cops Use AI, Arrest the Wrong Guy

Our Take

The Civil Rights Lawyer lays out a story that should sound familiar to anyone tracking Flock Safety's ALPR network: a machine makes a guess, a business or cop treats that guess as gospel, and an innocent person ends up in handcuffs. Swap 'facial recognition' for 'license plate reader' and the failure mode is identical — an algorithm spits out a confidence score, not a fact, and the humans downstream stop asking questions the moment a screen says 'match.'

This is the core problem with outsourcing suspicion to AI: these systems are optimized for speed and coverage, not accuracy in edge cases, and there's rarely a meaningful human check before consequences hit. Flock's marketing leans hard on the idea that its cameras are just quietly logging plates for later — but the same logic in this video applies. Once a system flags you, the burden shifts to you to prove your innocence, not the other way around.

Credit to The Civil Rights Lawyer for walking through how fast an automated hunch can turn into a real arrest with real consequences. If you want to see how dense this kind of surveillance infrastructure already is near you, check our camera map, and if you're ready to push back on it locally, our take-action guide is a good place to start.