Robert Dillon was arrested over a 93% facial recognition match that was wrong #Shorts

Our Take

A 93% match sounds like rock-solid proof, but as the case covered in this short from TrendingTrauma shows, it's really just a statistic dressed up as evidence. Robert Dillon's lawsuit alleges that number translated into real handcuffs, a real arrest record, and real harm — over a face the system got wrong. Facial recognition and ALPR systems both trade on the same dangerous illusion: that a percentage or a plate read equals ground truth, when in practice it's often the start of a chain of human assumptions that nobody bothers to double-check.

This is exactly the failure mode we track with Flock Safety's camera network. A 'hit' on a scan — whether it's a face or a license plate — frequently becomes probable cause in the eyes of police, even though the underlying tech is probabilistic, not definitive. Dillon's case is a facial recognition story, but the civil liberties problem is identical to what we document with ALPRs: outsourcing judgment to a black-box algorithm, then treating its output as unimpeachable fact in an arrest warrant or a traffic stop.

Thanks to TrendingTrauma for putting a name and a face to what too often stays abstract. If you want to see how dense this surveillance infrastructure already is in your own neighborhood, check our camera map, and if you're ready to push back on unchecked deployment, our take-action page has concrete steps.