|
Getting your Trinity Audio player ready...
|
Geneva police, in Ontario County, say they keep license plate reader data for 21 days. The department’s written policy doesn’t say so, however. Fairport police, in Monroe County, do have a written policy that states how long such data can be kept—up to six years, but the village says the department last used its system around 2019 and holds no data from it.
The gap between the Geneva and Fairport policies shows how much of the rulebook for plate readers is left to individual police departments: not just how long records are kept, but who can search them and who checks. This primer on the technology and policies governing its use draws on police policies, public records and published statements. It is part of the Genesee Signal’s repository of police surveillance technology records. (A follow-up story will cover the debate over who should set the limits.)
What a license plate reader records
An automated license plate reader, or ALPR, uses cameras and software to capture images of license plates and convert the characters into searchable data. Depending on the equipment, it may also capture an image of the vehicle and record the time and location of the reading. Law enforcement groups note that the cameras photograph cars and plates, not people.
Police compare plate reads against a “hot list,” which may include stolen vehicles, cars tied to wanted people or missing-person cases, and other plates entered for law enforcement purposes. A match gives an officer a reason to look closer. It does not establish that a crime has occurred or who was driving, and officers are expected to verify an alert before acting, because a misread plate or an outdated record can produce a false match.
Readers can be mounted on patrol cars or fixed at an intersection or roadway. When the records are kept and made searchable, investigators can look back at where a vehicle appeared over time. That is much of the technology’s investigative value, and it is why retention rules matter. A department that keeps records for a year can reconstruct a vehicle’s movements months after an incident; one that deletes them quickly may not have the information when a case develops later. Repeated reads can also show patterns of movement, including visits to sensitive places, though a vehicle’s presence somewhere doesn’t prove its owner was there.
The reader itself doesn’t identify anyone, law enforcement groups stress. The International Association of Chiefs of Police says officers can identify a registered owner only by querying a separate, restricted and audited state database, under the federal Driver’s Privacy Protection Act. That law restricts the release of personal information from state motor vehicle records, but it does not by itself govern how plate scans are collected, kept or shared. Critics say a long record of scans can reveal a great deal even without a name attached. Those questions depend on state and local law, department policy, vendor design and agreements between agencies.
Daniel Schwarz, senior privacy and technology strategist in the New York Civil Liberties Union’s policy department, says the invasiveness of a plate reader system comes down to a couple of things.
“When we think of ALPRs, the level of invasiveness comes down to two factors,” Schwarz says. “One is density and how many cameras there are in a particular area, and the other is retention.”
He says the risk grows when a department’s data is pooled into the large, centralized databases that vendors such as Flock and Vigilant operate.
“You end up with a mass surveillance database that can be used for all sorts of purposes and suddenly they’re no longer under local control,” he says. “Local departments don’t know who accesses their data.”
Two New York cases show how that plays out. In Troy, in Rensselaer County, the NYCLU says agencies across the country searched the city’s data more than 12 million times from April 2024 to March 2026. A report compiled from Troy police audit logs puts the figure at about 12.2 million, 94 percent of them from out-of-state agencies. The logs show who searched, not whether a search turned up a match. The report also shows Troy police turned off Flock’s National Lookup feature on March 18; outside searches, which ran above 700,000 in some sampled months, fell to about 38,600 in April, nearly all from New York agencies.
In Syracuse, police data appeared in nearly 4.4 million searches from June 2024, Central Current reported in July 2025, including 2,097 labeled immigration-related. Most were broad searches across many agencies’ databases, not aimed at Syracuse alone. Syracuse police said they inadvertently opted into Flock’s national sharing network, and a deputy chief told a council committee it happened with a click after other departments invited access. The Syracuse Common Council voted in March to revoke its Flock contract, according to the American Immigration Council’s policy tracker.
Schwarz says New York has no statewide privacy protections for these technologies. “It is a wild west,” he says, “where the government can do as it pleases and vendors can do as they please.”
Law enforcement groups view sharing differently. Thirteen national organizations, including the International Association of Chiefs of Police, the National Sheriffs’ Association and the National District Attorneys Association, said in principles issued Sept. 14 that law enforcement “must be able to responsibly share and query ALPR information across jurisdictions” because missing people and dangerous offenders don’t stop at jurisdictional boundaries. They say sharing must be limited to legitimate purposes and subject to controls and oversight, and that government agencies “must not sell access” to the data they collect.
Troy’s City Council has since passed a law limiting how long police there can keep the data and with whom they can share it.
The Geneva and Fairport police departments supplied their ALPR policies in response to public-records requests this fall. Both restrict use to legitimate law enforcement purposes and limit access. They differ sharply on retention and on what officers must record when they search.
Geneva’s General Order 640, effective Dec. 19, 2025, governs the department’s fixed Rekor Edge Max readers. It limits access to trained officers, requires them to visually verify an alert before a stop and to record how an alert was resolved, and assigns system administrators responsibility for access, collection, storage, retention and security. It classifies the data as for official use only and allows sharing for legitimate law enforcement purposes, with outside dissemination documented in an incident report or call-for-service record.
The order does not state a retention period, nor does it require a general query-by-query log. It leaves retention decisions to the chief or assigned administrators. Chief Ron Eveland said that the department keeps the data for 21 days, mirroring the city IT department’s retention period, and that Geneva shares its plate reader data with the Monroe County Crime Analysis Center. He said no other outside agencies have access. As of Oct. 5, he had not located the agreement governing that sharing.
Fairport’s General Order 514-18, effective June 12, 2018, takes a more stringent approach to documentation but keeps data much longer. It requires ALPR data to remain searchable and auditable for one year, then be archived for five more years before being destroyed in the ordinary course. Retention can be extended for ongoing investigations with written authorization from the district attorney’s office, which is also required to access archived data except when the law or a court order requires it.
For accountability, Fairport requires each search to include its reason, the identity of the authorized person conducting it and a case number when applicable, with query logs retained in the Vigilant database for audit. Access is limited to trained, authorized staff with unique credentials, and sharing with other agencies must be authorized by the department head or a designee, with the reason and the requester documented.
Fairport’s policy is clearly more specific. It also provides for keeping data far longer: six years in total, against the 21 days the Geneva chief described. The NYCLU is pushing a statewide bill, A.10808, that would require deletion within 48 hours, though only for data not being used, so the comparison isn’t exact. Geneva’s figure comes from the chief, not a written retention rule, and Fairport’s six years is what its policy allows, not a practice the village describes. In written answers, Village Manager Bryan White said the department last used the system around 2019, that the equipment can’t be located and that no data remains, though the policy is still in effect. The records do not show how the policy worked while the system was in use.
The policies show what the departments say their rules are. They don’t show how often either department searches its data, who outside the department can see it, or whether the rules are followed. Answering those questions takes query logs, audit results, retention settings, contracts and data-sharing agreements.
Schwarz says the public can rarely tell legitimate searches from illegitimate ones, even with audit data.
“I think it’s impossible to assess from the outside,” he says. For plate readers, “we’d want to see the locations of where those cameras are being deployed, on top of general policies, guidelines, and restrictions for officers.”
The positions below come from the companies’ websites and published contract terms. This section covers the two vendors named in this primer, Flock and Vigilant. Rekor, which supplies Geneva’s readers, is outside its scope.
Flock Safety says customers own and control their data. In August, the company cut its default retention from 30 days to seven and said it will make its Audit Assistance tool, which flags unusual searches, mandatory for law enforcement customers by the end of the year, according to The Next Web. Motorola Solutions, which owns Vigilant, says in its contract terms that agencies own their plate data and set their own retention, and that sharing is optional and logged for audit.
Geneva and Fairport show that departments can adopt very different rules for plate readers, and that a written policy doesn’t say how a system works in practice or whether it is still in use.
Next: Whether police administrators, city councils, or the state should set the rules governing ALPRs
Will Cleveland is a Rochester Beacon contributing writer and a former Democrat and Chronicle reporter. He writes about beer in the Finger Lakes region and Western New York at Cleveland Prost and publishes investigative pieces at The Genesee Signal.
The Beacon welcomes comments and letters from readers who adhere to our comment policy including use of their full, real name. See “Leave a Reply” below to discuss on this post. Comments of a general nature may be submitted to the Letters page by emailing [email protected].