NYPD's facial recognition landed the wrong man again, proving the danger of AI

Our Take

Doom Machine's latest breakdown of NYPD's facial recognition mess is worth your time because it's not an isolated glitch — it's a pattern. Wrongful arrests from face recognition keep happening, and the departments deploying this tech keep insisting the next version will be the reliable one. It never is. The people who pay for that overconfidence are almost always Black men who get hauled out of their homes in front of their families for a match an algorithm was never built to guarantee.

We cover Flock Safety's ALPR network because it runs on the same logic: trust the machine, skip the skepticism, expand the dragnet. Face recognition and license plate cameras are cousins in the same surveillance family — both marketed as neutral tools, both quietly wrong at scale, both stitched into police workflows with little public oversight or chance to challenge a bad match. When NYPD's system flags an innocent person, it's not just a face recognition problem, it's a preview of what happens when any biometric or location-tracking system gets treated as infallible.

If you want to see how far this kind of automated tracking has already spread near you, check our camera map, and if you're ready to push back on it locally, our take-action page has a place to start. Credit to Doom Machine for keeping this story in front of people — these failures don't stay newsworthy long unless someone keeps counting them.