Flock’s Cameras Log Every Passing Car. A Lawsuit Says the Safeguards Arrive Too Late

A former Georgia deputy allegedly searched one woman’s license plate 1,639 times over several months. Each search left a record. According to a proposed class action against Flock Group Inc., those records were detailed enough to reconstruct the activity later, but the platform did not interrupt it while it was happening.

That gap between recording misconduct and preventing it is at the center of the complaint. Plaintiff Jarrod Whitson says Flock’s cameras turn routine travel into a searchable location history, then make that history available through a platform used across thousands of law enforcement networks. The lawsuit alleges that Flock retained control over crucial search, sharing, logging, and access functions while leaving key anti-abuse protections optional, reactive, or absent.

A roadside sighting becomes a searchable itinerary

A single camera entry is narrow. It indicates that a vehicle passed a particular location at a particular time. The complaint alleges that Flock’s value comes from accumulating those entries, indexing them, and allowing authorized users to search records gathered by cameras their own departments did not install.

The cameras allegedly operate continuously and capture plate numbers, vehicle images, time, location, and visible characteristics such as make, color, or identifying marks. Repeated searches can arrange those sightings into a pattern. That pattern may expose regular trips to workplaces, homes, medical providers, places of worship, shelters, attorneys’ offices, or the homes of partners and relatives.

1. Routine captureA roadside camera photographs a passing vehicle and reads its plate without waiting for suspected wrongdoing.
2. Central indexingThe plate, image, place, time, and vehicle characteristics become searchable records in Flock’s hosted system.
3. Shared searchingDepending on permissions, a user may retrieve sightings collected by other agencies across a state or the country.
4. Recorded queryThe platform logs who searched, when, the stated reason, and the plate or search term, according to the filing.

Conceptual summary of the mechanism alleged in the complaint. Access and sharing can vary by agency settings and permissions.

Flock does not simply sell hardware, the complaint argues. It supplies the authentication, search, sharing, logging, and audit functions that make the network useful. The filing acknowledges that customer agencies decide who receives a login and how their available settings are configured. Its central contention is that Flock defines what those settings can do, writes the code, hosts the data, and retains contractual authority to restrict or suspend access.

The alleged misuse was repetitive, not momentary

The complaint cites an August 2026 journalistic tally of at least 50 officers who faced charges or allegations involving plate-reader systems. Flock’s platform reportedly appeared in 46 of those matters. In 26, the target was described as a wife, girlfriend, former partner, a former partner’s new companion, or another woman in whom an officer had taken an interest.

Those numbers come from reporting cited by the plaintiff, not findings made in this lawsuit. Arrests and criminal charges also do not establish guilt. Even with those limitations, the incidents presented in the complaint describe a recurring pattern: authorized credentials allegedly being redirected toward personal surveillance.

Georgia authorities also announced arrests in 2026 involving current and former law enforcement personnel accused of using Flock systems for non-law-enforcement purposes. The complaint says five former Albany officers faced counts ranging from one to eleven, while a Conyers dispatcher supervisor allegedly conducted more than 30 unrelated searches over roughly four months.

The plaintiff’s design argument relies on the volume. A mistyped justification might evade review once. Hundreds of searches against the same person are a detectable pattern, the filing argues. Flock’s logs allegedly preserved that pattern but did not automatically restrict, escalate, or stop the account.

The lawsuit is not only about rogue officers

The complaint identifies other pathways through which data or camera access allegedly moved beyond what a customer expected.

These accounts carry different evidentiary weight. The Illinois statement is presented as an official audit conclusion. The Mountain View and demonstration incidents are relayed through city records and reporting cited in the complaint. The court has not determined in this case whether they prove the plaintiff’s broader claims.

When billions of readings meet imperfect accuracy

The complaint cites two Flock performance figures: character recognition above 96 percent in good or optimal conditions, and plate-read accuracy of approximately 93 percent. It does not establish that those measures use identical definitions or testing conditions.

Using the lower figure and Flock’s reported volume, the plaintiff calculates that a 7 percent error rate would produce approximately 1.4 billion inaccurate readings each month. Using a 96 percent success rate would produce approximately 800 million. These are mathematical extrapolations, not an audited count of incorrect records. A bad reading also does not necessarily become a police alert or cause an encounter.

Roseville’s result is narrower and more concrete. According to the filing, the California city reviewed 1,427 Flock alerts from 2023 and 2024 that flagged vehicles as stolen or connected to felonies. It found a misread plate in 1,011 alerts. That alert sample is not directly comparable to all plate readings across Flock’s network.

The complaint says Flock responded by restating its greater-than-96-percent character-recognition figure under optimal conditions, attributing the results to Roseville, and asserting that performance had improved. Roseville reportedly disputed that characterization.

A correct plate can still support a wrong conclusion

The lawsuit separates machine error from human inference. A plate can be read correctly while the conclusion drawn from the vehicle’s presence is wrong. A car near a location does not identify the driver, establish a purpose, or prove a crime.

The complaint cites the case of a Colorado woman whose normal commute passed a home where a package had been stolen. Cameras accurately recorded her truck on that route. Police allegedly interpreted the recurring pattern as evidence and accused her of theft. She reportedly assembled GPS logs, vehicle records, and video to clear herself.

Other reporting cited by the plaintiff describes motorists being pulled over, handcuffed, jailed, or approached by armed officers after misread plates or faulty alerts. The complaint does not supply the underlying records for each incident, but it uses them to argue that an error in a location system can migrate quickly from a database into a physical encounter.

For people targeted by personal searches, the alleged harm is different. Repeated queries can disclose routines without alerting the person being watched. The subject does not receive notice when a search occurs, according to the filing, even though the platform retains an audit trail.

Whitson’s own record exposes an important unknown

Whitson, a Tennessee resident and law enforcement officer, says he routinely drove past Flock cameras and later learned through a public-records request that his plate data had been queried across 5,675 separate networks. The complaint does not say that this figure represents 5,675 individual searches.

It also does not identify who initiated the queries, why they occurred, when they occurred, or whether any were unauthorized. Those missing facts matter. They prevent the filing from showing, based on the supplied record alone, that Whitson personally experienced the kind of targeted misuse described elsewhere.

His broader claim is that capture itself caused a privacy injury because Flock allegedly recorded, retained, and made his movements searchable without notice, consent, or individualized suspicion. Whether that theory establishes a legally compensable injury for him or the proposed class has not been decided.

Flock’s reported answer: misuse is rare and controls exist

The complaint includes several elements of Flock’s public position. It says Chief Executive Officer Garrett Langley described abuse as exceptionally rare in an interview and discussed misuse as a trade-off against the number of investigations supported by the technology. The filing also acknowledges that Flock provides compliance tools and shortened its default search-data retention period from 30 days to seven.

Flock reportedly said future sales demonstrations would use more public cameras after the Dunwoody incident. In the Roseville dispute, it defended its recognition performance and attributed the city’s results to local implementation or conditions.

The plaintiff’s reply is that shorter retention does not prevent repeated searches during the remaining window, and that optional controls do not amount to default protection. The complaint says Flock has offered functions such as excluded search terms, case-number requirements, sharing controls, and automated monitoring for atypical patterns. It argues that these features demonstrate the technical feasibility of stronger preventive safeguards.

Six claims, no ruling yet

Whitson seeks to represent a nationwide class defined as every person whose vehicle was photographed and whose plate was read, recorded, and retained by a Flock camera during the proposed class period. Twenty billion monthly readings do not equal 20 billion people, and the complaint does not estimate the number of unique proposed class members.

The lawsuit asserts six claims: negligence per se under Georgia’s automated license plate recognition law, intrusion upon seclusion, negligence, negligent design, failure to warn, and intentional infliction of emotional distress. The plaintiff seeks damages, attorneys’ fees, and class certification.

The negligence per se claim argues that Flock violated a Georgia statute restricting access to captured plate data to law enforcement purposes. Georgia prosecutors have reportedly charged individual users under that law. Whether the statute also supports civil liability against Flock on the facts alleged is a question for the court, not an established conclusion in the filing.

The requested injunction would require preventive controls against prohibited access, real-time identification of abnormal or repetitive searching, and interruption of identified misuse. Those are remedies requested by the plaintiff. None has been ordered in the supplied record.

The core question is prevention, not paperwork

The complaint presents Flock’s audit trail as both evidence of accountability and evidence of its limits. Logs can identify who searched, what they searched for, and the reason they entered. That information can support an investigation after misuse is discovered.

What logs cannot do by themselves is protect a person during the 69th, 600th, or 1,639th alleged search. The lawsuit asks whether a company that built a national vehicle-search platform had a duty to recognize those patterns sooner and interrupt them. The plaintiff has not yet proved that case. The incidents assembled in the filing nevertheless expose the stakes of the design choice: a system can preserve a precise record of abuse while the person being tracked remains unaware that it is happening.