Wrongful Arrest Exposes Flawed Florida Police Face Recognition System
Original: Wrongful Arrest Exposes Failures in One of the Oldest Police Face-Recognition Tools in the US
Why This Matters
Highlights ongoing accuracy issues with police facial recognition technology
Robert Dillon, a 52-year-old Florida commercial crabber, was wrongfully arrested after FACES face recognition system returned a 93% match linking him to a child abduction attempt 300 miles from his home in Jacksonville Beach.
The ACLU filed a lawsuit against two Florida police departments after Robert Dillon was arrested based on a flawed face recognition match from FACES, one of America's oldest police facial recognition databases holding tens of millions of Florida photos. Despite living 300 miles away and never visiting Jacksonville Beach, Dillon was arrested after the system matched his face to surveillance footage of a man who approached a child at McDonald's. Police ignored evidence pointing away from Dillon, including that the suspect was a regular customer at the location and license plate readers showed Dillon's vehicles were never in the area. Dillon was held overnight, posted bond using his truck title, lost income during peak crab season, and his mugshot remained online for nearly a year.