In December of last year, an internal-affairs detective at the Milwaukee Police Department opened an investigation into one of the department's officers, accused of abusing Flock — the agency's automatic license-plate-recognition system. By February, the criminal complaint laid out details that were startling: Officer Jose Ayala had allegedly used Flock 124 times to look up the location of his romantic partner, and another 55 times to track the partner's ex-boyfriend.
The abuse was not an isolated case. Earlier this month, the very detective who had investigated Ayala was himself arrested for abusing Flock — he admitted to using the system to track two individuals, and to secretly planting a GPS device on one of their cars.
This is not a story about a few bad cops. Statistics compiled by the Justice Institute, a nonprofit public-interest law firm, show that across the country at least two dozen cases have involved officers resigning or being arrested for allegedly abusing Flock to stalk romantic partners — and the publicly reported cases are only a fraction of the total. Last month, five officers in Albany, Georgia were arrested and fired for abusing Flock; two officers in Greer, South Carolina also lost their jobs for the same reason.
A $7.5 Billion Surveillance Company
Flock Safety is the largest automatic license-plate-recognition (ALPR) company in the country, headquartered in Atlanta, working with more than 5,000 law-enforcement agencies, with a network that spans 49 states and comprises more than 120,000 cameras. Founded in 2017, the company secured a $275 million funding round in March of last year — led by major Silicon Valley investors including Andreessen Horowitz and Founders Fund (the firm backed by Peter Thiel) — that valued it at $7.5 billion.
The technology itself is not mysterious: cameras mounted on police vehicles or on fixed roadside poles photograph license plates and log them in a database. When a vehicle on a watch list (linked, for example, to a crime or to a missing-person case) passes by, the system issues an alert; officers can also directly query a plate to see where it has recently been; and different departments can share data with one another.
Law-enforcement officials stress its case-solving value: police last year used Flock cameras to track down the suspect in the Brown University mass shooting; in Milwaukee, ALPR helped close 24 cases in April and 26 in May, leading to arrests for domestic violence, robbery, and strangulation warrants. Flock claims it supported roughly 1 million investigations last year and helped locate more than 10,000 missing persons.
These numbers are real. But "10,000 missing persons" and "124 queries on a romantic partner" live inside the same system — and that is exactly the shape of the problem.
Where the Abuse Comes From
Look at the common thread across these cases: the abusers did not need any hacking technique, nor did they need to bypass permissions — they were simply the people authorized to use the system. Detective Ayala had access, so checking his partner's location was just "a few extra keystrokes"; the detective investigating him had the same access.
In other words, the security boundary of this system is not technological — it lies in human discipline. Flock itself is aware of this. CEO Garrett Langley has said the company rolled out an "audit-assist tool" last month that alerts auditors when "anomalous" usage is detected. His exact words: "This is a very powerful tool, and we don't want to slow down the 99.99% of officers who show up every day and want to do good things for their communities."
That sentence is worth reading twice. It acknowledges that the system will be abused, and then chooses to plug the gap with an "audit assist" — rather than designing the system to constrain the purposes for which queries can be made. The cost of abuse is borne by the people being stalked; the burden of remediation is placed on an auditing tool and on the professionalism of "the 99.99% of good cops."
The Governance Problem of a Surveillance Net
Placed in the broader picture, Flock represents a typical configuration of America's law-enforcement surveillance infrastructure: built by a private company, used by government agencies, with data shared across departments. With 120,000 cameras, whether you drive or not, whether or not you have a record, your plate's movements are inside this net — the point is not "whether you have been looked up," but "the person who wants to look you up can do so at any time."
The two dozen stalker cases tallied by the Justice Institute are only the ones that have surfaced. Flock's network covers 49 states and more than 5,000 law-enforcement agencies, and the system supports roughly 1 million investigations a year — at that scale, no one can put a number on the abuse that has not been discovered.
Langley's "audit-assist tool" is a step in the right direction, but it places the center of gravity of governance on "anomaly detection after the fact." The real watershed is this: when surveillance infrastructure grows large enough to monitor everyone, anti-abuse safeguards should be designed in upfront — whether a query needs a justification, how long the data is retained, where the boundary of cross-departmental sharing lies — rather than relying on post-hoc auditing and individual professional conduct.
When surveillance infrastructure grows large enough to monitor everyone, anti-abuse safeguards should be designed in upfront — whether a query needs a justification, how long the data is retained, where the boundary of cross-departmental sharing lies — rather than relying on post-hoc auditing and individual professional conduct. The configuration of private companies building, government agencies using, and data shared across departments makes "anyone who wants to look you up can do so at any time" the system's default — and every step forward by the counter-movement is, in essence, stripping a layer off that hidden surveillance.
From Whistleblowing to a Map — DeFlock, Lawsuits, and Legislation
Increment appended 2026-08-11: Two months on, the controversy around Flock has taken on a new shape: opponents are no longer merely filing complaints with police departments — they are now mapping the cameras' locations across the country.
A long-form investigation by The New York Times provides the latest coordinates: Flock has now installed roughly 120,000 cameras in every state except Alaska, with contracts covering 7,000 law-enforcement agencies and roughly 40 percent of all U.S. law-enforcement departments, while the company is valued at $8.4 billion. Texas is a microcosm of the expansion — 700 police departments there have signed contracts with Flock; the liberal city of Austin and the conservative town of Bandera (population 900) have both canceled their contracts, with at least one camera in Bandera vandalized; Los Angeles and Dayton, Ohio suspended cooperation to keep data out of the hands of immigration authorities, with Dayton even covering cameras with garbage bags; in Houston, Oakland, and Grand Junction there have been incidents of cameras being cut, spray-painted, and even shot.
The Crowdsourced Map: DeFlock
On the technical side of the counter-movement, there is a crowdsourced map. Software engineer Will Freeman, on a 2024 drive from the West Coast to Alabama, noticed cameras along the way and created the DeFlock app, inviting users to tag the locations of license-plate-recognition devices; downloads have now exceeded 350,000, and the map marks 125,000 ALPRs produced by Flock and other companies. Philip, a 27-year-old signal-tower technician, drives around the outskirts of Dallas adding cameras to the map one by one — his reasoning is plain: "So how am I supposed to fight back? Through education and data."
Lawsuits and Legislation: Putting Constitutionality on the Table
On the legal side of the counter-movement, lawsuits are beginning to challenge the constitutionality of the entire system. In May of this year, Freeman filed suit against the City of Boulder, arguing that the cameras constitute an unconstitutional "unauthorized mass surveillance network" — Boulder had installed 31 cameras and motor-vehicle thefts dropped by 34.5 percent, but the city's response asserts that "on public streets and sidewalks, there is no reasonable expectation of privacy." At the federal level, Republican Congressman Keith Self (a retired Army colonel) introduced a bill in July requiring federal agencies to obtain a warrant before querying state and local surveillance data; the Policing Project at NYU School of Law counts 13 states that have already passed legislation requiring audits of license-plate-recognition systems.
The opposition to Flock has completed a three-stage leap — from "individual whistleblowing" to "crowdsourced mapping + constitutional lawsuits + legislative audit demands." What makes the network genuinely difficult to sustain is not the abuse of individual officers, but the design itself — "anyone who wants to look you up can do so at any time" — being put on the table for review: once camera locations become public data, the hiddenness of surveillance is stripped away; once warrant requirements become law, the cost of abuse no longer rests solely on post-hoc auditing.