businessinsider.com web signal

Flock plate reader misreads keep flagging innocent drivers

TL;DR

  • Business Insider's ongoing review of police records and lawsuits has documented roughly a dozen instances of Flock plate misreads leading to wrongful stops.
  • An LAPD inspector general audit found about one in three Flock hot-list alerts reviewed were inaccurate, and Los Angeles let its Flock contract expire in July 2026.
  • Flock says its cameras accurately capture 93% of the roughly 20 billion license plates its readers scan each month.

Business Insider has spent this year working through police records, lawsuits and local news to build a picture of what happens when Flock Safety's automated license plate readers guess a character wrong. The pattern from that reporting is fairly consistent. A plate is misread, the system flags the wrong car as stolen or wanted, an officer treats the ping as fact, and someone who has done nothing gets pulled from their vehicle at gunpoint, sent to jail, or in one documented case attacked by a police dog.

The specific misreads look small on their own. A camera read a '7' as a '2' and the registered owner was transported to a hospital for injuries and held for hours before charges were dismissed. A separate camera read an 'O' as a '0' and two grandparents were detained at gunpoint in front of their three-year-old granddaughter. Flock's own claim is that its readers accurately capture 93% of the roughly 20 billion plates its cameras scan each month. That sounds tolerable as a marketing line, but even a modest error rate at that scale is a large number of wrong alerts landing on officers' screens.

The gap between vendor-quoted scan accuracy and department-level alert accuracy is where the story really lives. An LAPD inspector general audit that reviewed the hot-list alerts Flock actually sent to officers found roughly one in three flagged vehicles were not actually stolen, and Los Angeles let its three-year Flock contract expire in July 2026 on the back of that finding. Roseville, California is another department cited in the same reporting arc; it moved to Flock in 2021.

The honest caveat is that Flock has disputed the framing of many of these incidents in prior statements, arguing its cameras worked 'correctly' and that the fault lay with stale entries in the underlying stolen-vehicle databases rather than with its optical character recognition. What the reporting does not fully give you is whether departments train officers to independently verify a plate before treating a Flock ping as probable cause for a felony stop, which is arguably the load-bearing question.

The forward-looking part is that once one city auditor writes up a 'here is our real false-alert rate' report, others can copy the format. That is the fix worth watching for, more than any single lawsuit.

Shared on Bluesky by 2 AI experts