Flock Safety cameras: A breakdown of how automated license

Our Take

Credit to JarvisIT News for pulling back the curtain on something Flock Safety and its police partners rarely want to discuss in public: the algorithm behind the camera isn't infallible. A system marketed as a precision tool for catching stolen cars and wanted suspects is still, at its core, pattern-matching software making guesses about blurry plates, bad lighting, and look-alike vehicles. When that guesswork gets it wrong, it's not a glitch that stays in a server log — it's a real car pulled over, a real person facing an officer with a gun, based on a false positive.

That's the part of this story that deserves way more attention than it gets. Departments buy these systems on the promise of near-perfect accuracy, but 'near-perfect' still means thousands of towns generating thousands of chances for the software to flag an innocent driver as a threat. Multiply a double-digit error rate across a national network of cameras scanning every plate that passes, and you get a system that isn't just imperfect — it's structurally guaranteed to misidentify people at scale, with almost no public accounting for how often that happens or what happens to the people it happens to.

This is exactly why we track where these cameras are going up and push for basic transparency and oversight before communities say yes to them. If you want to see what's already deployed near you, check the map, and if you want to push back on a rollout in your town, our take-action page has the tools to start.