Police facial recognition tech led to wrongful arrest, lawsuit states

Our Take

Credit to the Tampa Bay Times for digging into a case that should worry anyone paying attention to automated policing tools, whether it's face recognition or license plate cameras. Robert Dillon spent time in jail because a computer told a detective he looked like a suspect, and apparently nobody double-checked before slapping on cuffs. That's not an edge case — it's the predictable outcome of building law enforcement workflows around algorithmic 'matches' that get treated as ground truth instead of a lead that needs actual verification.

What makes this especially relevant to our beat is the plumbing behind it: the Pinellas County Sheriff's Office runs this facial recognition system, apparently on behalf of agencies well outside its own jurisdiction. That's the exact same centralize-and-share model Flock Safety has built its business on with ALPR data — one agency's tech, quietly serving searches for departments hundreds of miles away, with little public accounting of who queried what, when, or why. When the underlying system misfires, the person left holding the bag isn't the vendor or the sheriff who manages the database. It's a guy who was just fishing and got arrested for a crime he had nothing to do with.

Wrongful arrests from algorithmic tips aren't rare anymore, they're a pattern — Detroit, New Orleans, and now Florida. If your local department is running Flock cameras or leaning on shared biometric databases with zero public audit trail, that's a story, not a footnote. Check our camera map to see what's already watching your community, and push your city council for real answers before your neighbor becomes the next Robert Dillon.