The scene: a Kohl’s parking lot in Plymouth, Minnesota. Multiple police vehicles box in a borrowed Range Rover on manufacturer plates. Officers take weapons-ready positions, treating the driver like a luxury car thief. That is exactly what happened to an automotive journalist after Flock Safety‘s AI-powered license plate reader network spent days tracking the vehicle across the state, flagging it as stolen on every camera it passed.
Neither the plate nor the vehicle was stolen. A Jaguar Land Rover dealership in Los Angeles had misplaced a plate during a photo shoot and reported it as lost. One clerical error, propagated through a nationwide surveillance network, turned a press loaner into a fugitive vehicle.
One Typo, One Nationwide Alert
A truncated plate number entered into Flock’s system created a false match against an entire class of manufacturer tags.
The misplaced plate read NJ 34 03 DTM. Someone entered it as “34 DTM” — dropping the middle digits entirely. JLR’s manufacturer plates use a format where those middle numbers appear in smaller print, and Flock’s optical character recognition reportedly failed to distinguish them. The cascade:
- The plate was misplaced at a photo shoot, not stolen
- The truncated entry matched every JLR media fleet plate formatted as “34 ## DTM”
- Plymouth police received repeated Flock alerts and tracked the journalist’s Range Rover for days before staging the ambush
- The responding officer noted that in Minneapolis, the stop would likely have involved drawn firearms
Flock’s own policy states officers must manually verify ALPR hits before taking action. Independent testing by IPVM, as reported by Business Insider, found the system “regularly misclassified license plate state, vehicle type, and make.” In Oak Park, Illinois, 40% of 25 Flock-triggered stops were mistakes caused by bad data or verification failures. Flock’s accuracy problems are structural, not isolated incidents — a pattern consistent with broader tech scandals in which systemic design failures go unaddressed.
Your Plates Are Being Watched. The Data Might Be Wrong.
Mounting evidence from multiple cities suggests Flock’s accuracy problems are structural, not isolated incidents.
Consider the credit report analogy: a single bad data entry blocks a mortgage application. Here, the consequence is armed officers surrounding a vehicle, not a declined loan. Mountain View, California suspended its entire Flock program after discovering federal agencies had accessed local camera data through a “nationwide search” setting enabled without city permission — essentially secretly tracking users by leaving location sharing on for every federal app without ever agreeing to it. A Toledo Police Department lawsuit alleges it is “commonly known” internally that “the Flock system is unreliable and often misreads license plates,” as reported by Business Insider.
The journalist’s situation resolved only after an on-scene call to Jaguar Land Rover confirmed the plate was never stolen. Yet while that single erroneous entry lived in Flock’s database, every JLR media fleet vehicle carrying a similar tag faced identical risk nationwide. The system performed exactly as designed. That structural failure is the detail demanding urgent policy scrutiny.
