Over a 60-day window in late 2025, automated license plate readers (ALPRs) operated by the Los Angeles Police Department generated 210.5 million plate reads and triggered 161 “stolen vehicle” alerts that officers later confirmed were false. According to a July 13, 2026 report by Jason Koebler at 404 Media summarizing the LAPD Office of the Inspector General’s (OIG) audit of August 1 to September 30, 2025, those 161 false alerts produced high-risk traffic stops on drivers whose vehicles were never stolen. The audit has now directly led the LAPD - the third-largest police department in the United States - to let its contract with Flock Safety, the AI-integrated ALPR vendor, expire.
The audit is the cleanest public read yet on what networked plate-reader infrastructure does to innocent drivers. It also makes concrete a question that has been theoretical for years: when the cameras and the hot lists they check against are both national, who pays for the inevitable false positives?
The audit, by the numbers
The OIG’s audit window covers August 1 through September 30, 2025 - a 60-day period. The numbers reported by 404 Media are LAPD’s own:
- 210.5 million license plate reads generated by LAPD cameras.
- 161 alerts acknowledged by officers as accurate plate matches, where “subsequent investigations determined the vehicles were not stolen.”
- 5,911 different license plates tracked during the period.
- 4,575 of those tracked plates - roughly 77% - led to no action.
- 337 stolen vehicles recovered.
- 74 arrests linked to ALPR data.
Those figures are striking on their own, but the 161 number is what matters. Each of those 161 alerts was treated as a “high-risk stop” under LAPD policy - the kind that pulls in backup, requests air support, dispatches a supervisor, and ends with the driver ordered out of the vehicle. Per the audit summary, those protocols are what get applied to an ALPR hit flagged as stolen.
The OIG’s broader conclusion, quoted by 404 Media, is that ALPR inaccuracies “can affect individual liberty interests, erode public trust, and potentially create substantial legal and financial liability concerns.” The OIG recommended suspending deployment of any new ALPR cameras, halting new ALPR-related contracts pending public input, broadening the vendor reassessment, and strengthening oversight of ALPR data access.
Why the network breaks
LAPD’s roughly 2,000 ALPR cameras run on hardware from Motorola, Flock, and Axon. Department policy says officers should attempt to verify the accuracy of an ALPR alert before initiating a stop; per the OIG audit summary, that verification often does not happen. Dispatch is built around the alert and the verification step gets compressed or skipped.
The data itself is networked, which is where the false-positive problem compounds. The LAPD accesses Flock and Axon data through Flock’s backend via an Axon-Flock data sharing partnership. That means a plate that one agency entered into a hot list in another jurisdiction can generate a hit in Los Angeles, and the lag between “plate recovered” and “plate cleared from hot list” is exactly the kind of timing error that produces an innocent stop. As 404 Media reports, LAPD’s public line is that stolen flags “generally result from the timing of record updates outside of the Department’s control, such as delays by another jurisdiction or a vehicle owner in clearing a plate from a Hot List after a vehicle has been recovered or is no longer wanted.” That is the agency acknowledging that the integrity of its own enforcement depends on records maintained by someone else.
The wider ALPR ecosystem has been warning about this for years. An EFF Deeplinks analysis published May 26, 2026 by Dave Maass and Rindala Alajaji cataloged ALPR searches run for reasons as varied as school residency verification, employment background checks, and noise complaints - including a Georgia school district where residency checks made up more than half of all ALPR searches in a 14-month window, with queries spanning 5,800-plus networks nationwide. The LAPD OIG audit is the first inspector-general audit that quantifies the false-positive cost on the enforcement side.
What “161 stops” looks like in the real world
The 161 number is not abstract. 404 Media cites two cases tied to the same kind of networked ALPR error LAPD was auditing.
Joel Feder, an automotive journalist with The Drive, wrote that he was test-driving a $155,000 Range Rover press vehicle in Minnesota when a Flock camera misread his plate as a different plate reported lost, which the system interpreted as a stolen-vehicle hit. He was surrounded by four police cruisers with his wife in the passenger seat and held at gunpoint - “hands on their guns,” in Feder’s words - before the stop resolved peacefully. The error originated from a Los Angeles Jaguar Land Rover dealership that had entered the plate as stolen; four other vehicles with similar plates were being tracked in Minnesota that week.
In a separate case cited by 404 Media, a 23-year-old woman was jailed for 13 days after police searched for a black Dodge Durango tied to a fatal hit-and-run. The ALPR hit produced an enforcement action that took nearly two weeks to unwind.
These are the stops the LAPD OIG audited and counted. LAPD CIO Dean Gialamas, as quoted by Futurism on July 13, 2026, confirmed the contract decision: “This contract is not being renewed because of serious concerns around civil liberties and civil rights issues, particularly around privacy and the data that is being collected from these cameras.”
What This Means
The LAPD OIG audit is the first public, inspector-general-level quantification of networked ALPR false-positive cost in a major U.S. city. Three implications follow.
First, hot-list hygiene is an enforcement-safety problem, not an administrative one. If 161 stops in 60 days were false in a single jurisdiction, the multi-jurisdiction hot-list ecosystem that Flock and similar vendors sell produces thousands of innocent high-risk stops nationwide. The LAPD’s defense - that record updates lag in other jurisdictions - is an admission that the system’s accuracy depends on practices the LAPD cannot audit.
Second, vendor consolidation is the structural problem. The same audit names three vendors (Motorola, Flock, Axon) and a data-sharing partnership (Axon-Flock via Flock backend) for the LAPD’s access layer. When the same company brokers queries across departments, a single record error propagates across the network - the same pattern our prior surveillance coverage tracked inside DHS deployments. Audits like the OIG’s are the only mechanism that surfaces the propagation; vendor self-reporting does not.
Third, the LAPD’s contract decision is the model. Gialamas said LAPD will not enter new contracts until a full audit is complete - tying procurement to inspector-general findings rather than to vendor promises. Other departments running networked ALPR programs - the EFF’s analysis documents searches by Buford City Schools, Delhi Township PD, Ridgeland PD, and dozens of others - do not yet have the equivalent of the LAPD OIG audit. Watch for any of those jurisdictions to publish comparable numbers.
The Bottom Line
Over 60 days in late 2025, LAPD’s roughly 2,000 ALPR cameras generated 210.5 million plate reads and produced 161 false “stolen vehicle” high-risk stops. The OIG audit led directly to LAPD letting its Flock Safety contract expire - making LAPD, the third-largest U.S. police department, the highest-profile department to drop the vendor over data-integrity and civil-liberties concerns to date. The audit is the first hard quantification of the false-positive cost of networked ALPR enforcement, and it sets the bar every other major-city ALPR program will now be measured against.