AI License Plate Cameras Are Getting Innocent People Arrested

Our Take

The Institute for Justice is doing what journalists and lawmakers too often won't: counting the bodies. Their reporting lays out case after case of drivers who did nothing wrong getting boxed in by patrol cars, pulled from their vehicles, or detained at gunpoint because a license plate reader misread a character, mismatched a state, or flagged a car that looked kind of similar to a car on a hot list. These aren't hypothetical failure modes civil liberties groups warn about in the abstract — they're real people who had guns pointed at them because a machine made a guess and an officer trusted it without verifying.

This is the part of the ALPR debate that Flock and its law enforcement customers would rather you not dwell on: the system is sold as precision policing, but it runs on probabilistic pattern-matching that gets it wrong often enough to produce a growing casualty list. When a network is scanning every plate on every street, even a small error rate turns into a lot of wrecked afternoons, terrified families, and, in the worst cases, split-second decisions made by officers who've been primed to expect a dangerous suspect. Nobody voted to install a system where a software glitch can escalate into a felony stop.

Credit to the Institute for Justice for compiling these cases into one place and putting a number on what's usually dismissed as a rare mistake. If you want to see whether this dragnet has reached your own neighborhood, check our camera map, and if you'd rather not wait to become the next case study, our take-action page has concrete steps for pushing back at the local level.