How Accurate Are Flock's AI-Powered License-Reading Cameras?
Futurism reports: 404 Media revealed that a July 10 audit by the Los Angeles Police Department Office of the Inspector General caught the department's ALPR cameras generating 161 false stolen-vehicle alerts in just two months — each one ending with officers pulling over an innocent driver. Factoring in 337 alerts which "resulted in the recovery of stolen vehicles," the LAPD's cameras carry an…
404 Media's July 10 audit revealed that LAPD's ALPR cameras produced 161 false stolen-vehicle alerts in just two months, leading officers to pull over innocent drivers. Of the 337 alerts resulting in recovered stolen vehicles, each represented a one-in-three chance of pulling an innocent person over. The issue stems from the FBI's outdated National Crime Information Center (NCIC) database, which Flock's CEO Garrett Langley criticized for lacking feedback loop mechanisms.
Local agencies, like Atlanta, can suppress false alerts for a year, but the FBI should modernize the system by allowing agencies to flag invalid tags. In a worst-case scenario, a Sacramento police officer reported six false Flock alerts for the same vehicle, incorrectly stating it had been stolen or used in a felony crime. The Roseville Police Department found that Flock sent 1,427 alerts in 2023 and 2024, with 71% of them incorrectly reading license plates.
Factors contributing to the misreads included older hardware, placement of cameras, and specific deployment configurations. One incident resulted in a driver being mauled by a police dog due to a misread license plate. Flock acknowledged some improvement in Roseville's camera performance but disputed the police department's claims.
Misreads have also occurred in Toledo, Ohio, where a driver lost his job and home after a misread license plate incident. Flock's software has struggled with reading plates correctly, misclassifying vehicles, and struggling with similar characters.
Written by urgent.news from Slashdot's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
