Flock Cameras in Nassau County: What Long Island Drivers Need to Know About License Plate Readers

There are 47 documented automated license plate reader cameras in the Village of Hempstead. There are 46 in Old Westbury, a village of roughly 4,700 people. Across New York State, community trackers have mapped more than 4,600 of these cameras in 234 municipalities, plus another 1,600 in unincorporated county areas, operated by roughly 60 agencies.

Most Long Island drivers have never noticed one. They are small, solar-powered, mounted on slim poles at intersections, on parkway ramps, at the entrances to villages and shopping centers and condominium complexes. They do not flash. They do not issue tickets. They photograph the back of your car, read the plate, log the time and the location, and add one more dot to a map of where your vehicle has been.

If you drive on Long Island, you are in that database. Not because anyone suspects you of anything — because you drove past a pole.

I practice criminal and DWI defense in Hempstead, and over the last two years the phrase “license plate reader” has gone from an occasional footnote in a police report to a routine part of how cases begin. An officer gets an alert. A plate comes back with a suspended registration, an active warrant, a scofflaw flag. A stop follows. Then come the observations — odor of alcohol, glassy eyes, slurred speech — and a DWI arrest that started with a camera nobody was looking at.

This article explains what these cameras are, how heavily they are used in Nassau and Suffolk, what New York law actually says about them (less than you would hope), and what a defense lawyer should be demanding when one of them is the reason you were pulled over.

What a Flock Camera Actually Is

Flock Safety is the best-known vendor, and its name has become shorthand the way “Xerox” once stood in for photocopying. Other companies — Genetec, Axon, Motorola — sell comparable systems, and New York agencies use all of them. The underlying technology is the same.

An automated license plate reader, or ALPR, is a camera plus software. As a vehicle passes, the camera captures an image. Optical character recognition converts the plate into text. The system records the plate number, a timestamp, GPS coordinates, and usually a photograph of the vehicle itself — make, model, color, and what the industry calls “vehicle fingerprint” details like roof racks, bumper stickers, dents, and window decals.

That record goes into a searchable database. Two things happen with it.

Real-time alerting. The plate is checked against “hot lists” — stolen vehicle files, wanted person files, DMV suspension and revocation data, Amber and Silver Alerts, and lists individual agencies build themselves. A match generates an alert to officers in the area, often within seconds.

Retrospective search. This is the part most people do not understand, and the part that matters constitutionally. The database retains every plate, not just the ones that hit. An officer can go back later and ask where a specific plate has been over the past days, weeks, or months. Depending on the agency’s retention setting and its network sharing agreements, that search can span a single village or thousands of jurisdictions across the country.

The first function is a modern version of an officer running a plate. The second is something categorically different: a searchable, retroactive record of the movements of every driver who happened to pass a camera, held whether or not anyone was ever suspected of anything.

How Many Are on Long Island

The honest answer is that nobody knows exactly, because New York does not require agencies to disclose it.

What we have is a patchwork. Community-maintained mapping projects like DeFlock, combined with public procurement records, document roughly 4,660 cameras across 234 New York municipalities as of late August 2026. Nassau County villages appear repeatedly in that data: Hempstead at 47, Old Westbury at 46, Glen Cove at 21, and dozens of smaller villages running anywhere from one to a dozen. Suffolk shows a thinner but real footprint — East Hampton at 15, Quogue at 9, Southampton at 8.

These are documented cameras. They are a floor, not a ceiling. Volunteer mappers find what they can see from a public road, and a camera nobody has photographed is a camera that does not appear on the map. Private deployments — homeowners’ associations, shopping centers, apartment complexes, self-storage facilities — often feed into the same network and are almost entirely invisible to the public.

The picture upstate suggests how quickly this scales. In Erie and Niagara counties, investigative reporting documented roughly 395 readers operated by 20 agencies, with Buffalo alone spending $366,000 on a 14-intersection system and Amherst moving to a $300,000 contract for 24 cameras. Smaller towns pay smaller sums — Hamburg runs four cameras for about $12,000 a year — which is precisely why the technology spreads so easily. A village board can approve it out of a line item without a public hearing.

How Often Nassau and Suffolk Police Actually Search

This is where the numbers become concrete.

In June 2026 alone, according to audit data published through Flock’s own transparency portal:

That is roughly 2,800 database queries in a single month by three Long Island agencies. The stated reasons ranged from ordinary traffic violations to suspected terrorism.

For comparison, in Western New York, 20 police departments collectively ran more than 218,000 Flock network searches between January 2025 and June 2026. Some individual departments there share data with 170 outside agencies and can search across 600 or more networks. One campus police department was documented as having access to 982 jurisdictions nationwide.

Those sharing figures are the part clients find hardest to believe. A camera in a Nassau village is not just a Nassau village camera. Depending on the settings the agency chose, it may be a node in a network searchable by police departments in states that have never heard of Hempstead.

The Nassau County “Surveillance Wall”

In late November 2025, Nassau County Executive Bruce Blakeman announced in a Fox News interview that the county would install a line of surveillance cameras along the Nassau–New York City border. In his words: “We are installing technology along the border of New York City that will read license plates, that will have facial recognition, that will have video cameras.”

The stated justification was the incoming New York City mayoral administration and what Blakeman described as its “radical policies.” New York State Democratic Party Chair Jay Jacobs called the plan “high-tech theatrics.”

What the announcement did not include: how many cameras, at what cost, from which vendor, with what retention period, under what use policy, subject to what audit, and with which outside agencies the data would be shared. The Nassau County Police Department declined to comment at the time.

I raise this not to make a political point but a practical one. That is the entire disclosure. In New York, a county can announce a network combining license plate readers with facial recognition, decline to specify a single operational parameter, and be under no statutory obligation to say more. That is the legal environment Long Island drivers and Long Island defendants are operating in.

The Misuse Problem — and What New York Has Not Told Us

Nationally, the record on ALPR abuse is now extensive and it is not disputed.

A Washington Post investigation published August 3, 2026, based on an analysis of police and court records nationwide, documented at least 50 cases in which officers used Flock’s network to track romantic partners, ex-spouses, and women they were personally interested in. The Institute for Justice maintains a parallel running tally that identifies more than fifty incidents across at least twenty states.

The individual cases are worth reading because of the volume involved. A Texas officer ran an ex-partner’s plate more than 2,000 times over eleven months. A Georgia deputy ran a woman’s plate 1,639 times. A Florida officer ran his estranged wife’s vehicle more than 700 times, entering false justifications each time. A Louisiana deputy searched an ex-fiancée’s plate thousands of times over several years.

Note what those cases have in common: every one of them was discovered through the system’s own search logs. The audit trail is what caught them. It is also what nobody was reviewing until someone thought to look.

Here is the New York-specific point, and I want to be precise about it. As of this writing, the Institute for Justice tracker contains no documented incidents involving New York State law enforcement agencies.

There are two possible explanations. One is that New York officers have not misused these systems. The other is that nobody in New York has systematically audited the logs and published what they found.

New York has no statute requiring agencies to audit ALPR searches, no statute requiring them to publish audit results, no statute penalizing misuse specifically, and no statewide body reviewing the query logs. The Florida, Georgia, and Illinois cases surfaced because journalists, defense attorneys, and inspectors general obtained and read search logs. In New York, that work has largely not been done.

What New York Law Actually Says

Very little, and that is the central problem.

There is no New York ALPR statute

Florida has statutes addressing plate reader retention and confidentiality. Illinois and Utah have data-sharing and retention frameworks. California has a dedicated ALPR chapter in its civil code.

New York has none of this. There is no statute in the Vehicle and Traffic Law, the Criminal Procedure Law, or the Public Officers Law that specifically governs how long ALPR data may be retained, who may query it, what justification must be recorded, whether logs must be audited, or with which agencies the data may be shared.

What exists is a 2011 document from the New York State Division of Criminal Justice Services titled “Suggested Guidelines: Operation of License Plate Readers.” The operative word is suggested. It is guidance. It creates no rights, imposes no duty, and carries no penalty. It also predates the current generation of networked, AI-assisted, nationally searchable systems by fifteen years.

In practice this means retention periods, use policies, and sharing agreements on Long Island are set by individual departments, and often by the vendor’s default settings, which many agencies never change.

People v. Bushey: running a plate is not a search

The controlling New York authority on the front end is People v. Bushey, 29 NY3d 158 (2017).

In Bushey, a SUNY Buffalo officer ran a lawfully driven vehicle’s plate through the DMV database with no suspicion whatsoever. The plate came back with a suspended registration for unpaid parking tickets. The officer stopped the car, discovered a suspended license, and made a DWI arrest.

The Court of Appeals affirmed. Chief Judge DiFiore wrote that a driver has no reasonable expectation of privacy in the publicly displayed plate or in the DMV information connected to it, so running the plate is not a search at all and requires neither reasonable suspicion nor probable cause. The Court acknowledged “the possibilities of database error and police officer abuse” but held that those risks “do not create a legitimate expectation of privacy where none existed before.”

If you have been arrested for DWI after a plate-reader stop in Nassau County, Bushey is the case the prosecution will cite, and on the narrow question of whether the officer could run the plate, it is going to win.

But read what Bushey actually decided. It addressed a check of a plate against the DMV database for current registration and license status. It did not address a retrospective query of an aggregated, multi-jurisdictional location database. It did not address months of stored movement history. It did not address AI pattern analysis. The Court decided the case in front of it, which involved looking up whether a car was legally on the road.

The gap between “an officer may check whether your registration is valid” and “an officer may retrieve six months of your movements from a nationwide network without any suspicion” is enormous, and Bushey does not close it.

People v. Weaver: New York protects location privacy more than the Fourth Amendment does

This is the case that matters most for New York defendants, and it is a genuine advantage of litigating here.

In People v. Weaver, 12 NY3d 433 (2009), the Court of Appeals held that prolonged GPS tracking of a vehicle constitutes a search requiring a warrant under Article I, Section 12 of the New York State Constitution — and it reached that holding three years before the United States Supreme Court decided United States v. Jones.

The Court’s language is directly applicable to plate reader networks. It described “the continuous GPS surveillance and recording by law enforcement authorities of the defendant’s every automotive movement” as something that “cannot be described except as a search of constitutional dimension and consequence.” It emphasized what aggregated movement data reveals: trips to the psychiatrist, the abortion clinic, the union hall, the mosque, the criminal defense attorney’s office. It declined to follow the more permissive federal approach of United States v. Knotts.

Weaver is a state constitutional holding. New York’s Article I, Section 12 is textually identical to the Fourth Amendment, but the Court of Appeals has repeatedly held that it may be interpreted to provide greater protection — and in Weaver it did exactly that for prolonged location tracking.

The argument for a New York defendant is straightforward. A months-long retrospective query of a networked plate reader database produces the same kind of aggregated movement record that Weaver held was constitutionally protected. The mechanism differs — fixed cameras capturing many vehicles rather than one device attached to one car — but the product handed to the officer is the same: a map of where a specific person went, assembled after the fact, without a warrant.

No New York appellate court has squarely decided this question. That is not a weakness. It means the issue is open, and it means the defendant who raises it properly and preserves it is the one who makes the law.

Carpenter and Chatrie: the federal trend line

Two United States Supreme Court decisions frame the federal side.

Carpenter v. United States, 585 U.S. 296 (2018), held that acquiring historical cell-site location information is a Fourth Amendment search requiring a warrant, notwithstanding the third-party doctrine, because of what comprehensive location records reveal about a person’s life.

Chatrie v. United States, decided June 30, 2026, extended that reasoning. In a 5-4 opinion by Justice Kagan, the Court held that obtaining location history from Google through a geofence warrant is a Fourth Amendment search subject to judicial oversight. The majority rejected the government’s argument that opt-in location services made the data voluntarily shared, calling that consent “somewhat illusory.” It emphasized that even short windows of location data reveal visits to “the psychiatrist, the plastic surgeon, the abortion clinic,” and it declined to draw workable temporal lines, warning of a slippery slope.

Chatrie does not mention license plate readers. But the reasoning transfers. Legal scholars have argued since the decision that ALPR networks share the two features the Court found decisive: they are retrospective and they are indiscriminate. Michael Soyfer of the Institute for Justice has argued that courts should evaluate “what was in the database and not just what police happened to access at a point in time.” Professor Andrew Guthrie Ferguson has made the related point that “modern ALPRs are just the connecting point to a much larger system of personally revealing information stored in police and connected public databases.”

Flock has publicly disputed the analogy, arguing that its cameras capture “point-in-time images of vehicles in public view” rather than data from a personal device. That is a real distinction on the front end. It is a much weaker one on the back end, where a single query returns a compiled movement history.

The federal Circuits are not uniform. In United States v. Porter, No. 25-60163 (5th Cir. Mar. 17, 2026), the Fifth Circuit held that ALPR use is not a Fourth Amendment search, reasoning that a motorist “has no reasonable expectation of privacy in his movements from one place to another” and distinguishing plate readers from cell-site data because they capture only periodic snapshots rather than continuous tracking. A Virginia state judge similarly upheld Norfolk’s system.

Two things about Porter are worth noting. It was decided on March 17, 2026 — three and a half months before Chatrie, without the benefit of the Supreme Court’s most recent statement on retrospective location data. And its reasoning rests on the same Knotts premise that the New York Court of Appeals expressly declined to follow in Weaver. A Nassau County court is not bound by the Fifth Circuit, and Weaver gives it independent state grounds to reach a different result.

The Pending Legislation

New York State Senate Bill S9890, sponsored by Senator Michelle Hinchey, would create the framework New York currently lacks. As of April 13, 2026, it sits in the Senate Committee on Consumer Protection.

Its main provisions:

• Use limits. ALPR systems could only be operated for hot-list matching on serious crimes, parking enforcement, secured-area access control, toll collection, and vehicle and traffic law enforcement.
• 48-hour deletion. Captured plate data would have to be deleted within 48 hours, with carve-outs for evidence preservation, toll matters, criminal defense requests, and valid court orders.
• Three-year query logs. Agencies would have to maintain logs recording the entity and individual making each query, the date and time, and the scope.
• Sharing restrictions. No sharing except pursuant to a federal or state judicial warrant, to criminal defense counsel, in discovery, or through FOIL with redactions.
• Enforcement. A private right of action for damages up to $1,000 and Attorney General authority to seek injunctive relief.

Note the defense-side provisions. The bill expressly contemplates preservation for criminal defense requests and disclosure to defense counsel. Whatever happens to it legislatively, it reflects a recognized understanding that this data is as often exculpatory as inculpatory.

A committee posting is not a law. But it is worth knowing that a 48-hour deletion rule is on the table, because current retention on many Long Island systems is measured in months.

The AI Layer

In August 2026, Wired reported that Flock had been testing an AI tool called OS Investigate, previously known internally as Nightshift, with law enforcement partners across roughly 6,000 communities.

According to that reporting, the tool lets officers query the network through a chatbot interface and cross-reference results against police files, 911 logs, and commercial identity databases. One prewritten prompt reportedly reads along the lines of: find me witnesses based on vehicles most seen in [neighborhood] during [last 14 days] during [daily timeframe].

Read that prompt carefully. It requires no plate. It requires no crime. It requires no suspect. It asks the system to identify people by their movement patterns in an area over time.

A former officer quoted in the reporting called it “completely insane.” ACLU attorney Chad Marlow warned it would let officers “go fishing for crimes.” Flock acknowledged testing the product and said current capabilities “may not reflect” the final offering, and that it is distinct from its plate reader technology.

Separately, in August 2026, following the wave of stalking revelations, Flock announced new auditing and accountability controls for its law enforcement customers.

For defense purposes the significance is this: if the tool your client’s case came from is materially more capable than “a camera that reads plates,” the defense is entitled to know what it actually did. Which brings us to discovery.

If a Plate Reader Started Your Case

New York’s discovery statute is a genuine advantage here, and it is underused in ALPR cases.

CPL Article 245 requires automatic disclosure, on a statutory clock, with a presumption in favor of disclosure. CPL 245.20(7) states plainly that “there shall be a presumption in favor of disclosure” in interpreting the automatic discovery provisions. Under CPL 245.10, the People’s initial obligation runs 20 calendar days after arraignment for a defendant in custody and 35 days for a defendant at liberty.

You do not have to persuade a judge that you are entitled to this material. The statute presumes it. The following subsections of CPL 245.20(1) do the work:

(g) — Electronic recordings. All tapes and other electronic recordings connected to the alleged incident. The plate reader capture itself is an electronic recording: the image, the plate read, the timestamp, the GPS coordinates.

(u) — Electronically created or stored information. All electronically created or stored information obtained by or on behalf of law enforcement. This is the provision that reaches the database query and its results, not just the single photograph the officer chose to print.

(j) — Reports, records, data, calculations concerning scientific tests, experiments, or comparisons. Optical character recognition is a comparison process performed by software with a measurable error rate. Plate misreads are well documented — B read as 8, O as 0, obscured or damaged plates, out-of-state plate formats. If the entire basis for the stop was a machine’s reading of six or seven characters, the reliability of that reading is squarely at issue.

(f) — Expert opinion evidence. If the People intend to rely on how the system works, they owe you the witness, the CV, and the proficiency testing.

(k)(iv) and (k)(vii) — Impeachment material and material providing a basis for a suppression motion. This is the subsection that reaches the audit trail. Evidence that the officer who ran the query has a history of improper searches impeaches that officer. Evidence that the query lacked a legitimate justification supports suppression.

The specific demands to make

Do not ask for “any license plate reader records.” Ask for these, by name:

1. The full audit log for every query of the ALPR network involving the defendant’s plate or vehicle, including the querying agency, the individual user ID, the date and time of each query, the stated reason or case number entered, and the date range searched.
2. All hits on the defendant’s plate within the retention window, not only the hit the officer relied on. A full read history establishes the scope of what the government actually collected and can independently establish that your client’s vehicle was elsewhere at a relevant time.
3. The agency’s written ALPR use policy, retention policy, and any memoranda of understanding or network sharing agreements governing which outside agencies could access the data.
4. The hot list source and entry. If the stop was based on an alert, what list generated it, when was the entry made, when was it last validated, and was it still accurate at the moment of the stop? A stale suspension that was cured weeks earlier is a live suppression issue.
5. The raw image and read confidence data. The system typically records a confidence score for the OCR read. A low-confidence read that an officer treated as certain is impeachment material.
6. Any AI or analytics tool used, including the name of the tool, the prompt or query submitted, and the output returned.
7. Whether a warrant was obtained, and if not, what the querying officer entered as the justification.

That last question is often the whole case. If the officer entered a case number that does not correspond to any existing investigation, or entered a generic justification like “investigation” for a retrospective search of your client’s movements, you have both a suppression argument under Weaver and a serious credibility problem for the People’s witness.

Preservation matters immediately

Retention windows on these systems are commonly 30 days. Some are shorter. If you wait for the ordinary discovery timeline, the exculpatory read history may be gone before anyone asks for it.

Send a written preservation demand to the police agency and the District Attorney’s office the day you are retained. Identify the plate, the date range, the specific system, and the fact that the material is discoverable under CPL 245.20(1). If the data is destroyed after a preservation demand, that destruction is itself litigable.

How to Find Out If You Have Been Searched

Two tools exist, and both are worth knowing about.

Flock’s transparency portal and the “Have I Been Flocked” project. Flock publishes audit data by agency, and third-party projects have made it searchable. You can review how many searches an agency ran, the reasons officers entered, and in some jurisdictions whether a specific plate appears in the logs. This is how the June 2026 Nassau and Suffolk search totals cited above became public.

DeFlock. A crowdsourced map of reported ALPR camera locations, including New York. It is incomplete by design — it depends on volunteers photographing cameras — but it will show you a great deal about your own neighborhood.

FOIL. New York’s Freedom of Information Law, Public Officers Law Article 6, reaches agency ALPR policies, contracts, procurement records, and in many cases audit logs. Agencies frequently resist on law enforcement exemption grounds, but the exemption is not automatic and policies and contracts are rarely properly withheld. A FOIL request is slower than a discovery demand but it is available to anyone, not just a defendant.

Frequently Asked Questions

Are Flock cameras legal in New York? Yes. There is no state statute prohibiting them, and no statute specifically regulating them either. Under People v. Bushey, running a plate against DMV records is not a search and requires no suspicion. What remains unsettled is whether a retrospective search of a networked movement database requires a warrant — an open question under both Carpenter/Chatrie federally and People v. Weaver under the New York Constitution.

Can a plate reader alone justify pulling me over? It can be the basis for a stop if it produces a lawful ground — a suspended registration, a revoked license, an active warrant, a stolen vehicle report. It cannot substitute for that ground. If the alert was based on stale, erroneous, or misread data, the stop and everything that followed are challengeable.

Can these cameras issue me a ticket? Not by themselves. ALPRs are distinct from red-light and school-zone speed cameras, which are separately authorized under the Vehicle and Traffic Law and do generate notices of liability. A plate reader generates an alert to an officer, not a ticket in the mail.

How long is my data kept? It depends entirely on the agency, because New York does not set a limit. Thirty days is a common default. Some agencies keep it longer. Pending Senate Bill S9890 would require deletion within 48 hours.

Do these cameras use facial recognition? Standard ALPR captures plates and vehicle characteristics, not faces. However, Nassau County’s announced border camera network expressly includes facial recognition as a separate component, and vendors have been expanding into appearance-based search. These are converging technologies.

Can ICE or federal agencies see New York plate reader data? Potentially, depending on the agency’s network sharing settings. Documented incidents in other states have shown federal agencies obtaining access through local departments’ network sharing, sometimes without the local department’s knowledge. New York agencies vary widely in how they have configured sharing, and most have not disclosed their settings publicly.

I was arrested for DWI after a plate reader stop. Does any of this help me? It can. The question is whether the stop was lawful at its inception and whether the data behind it was accurate. It is also the case that the same system that put you at the scene may have read history establishing where you actually were and when — which can matter for timing of consumption, route, and the reliability of an officer’s account.

How do I find out if my plate has been searched? Start with the Flock transparency portal and the “Have I Been Flocked” tool. If you have a pending case, your attorney should demand the audit log through CPL 245.20 discovery, which is faster and broader than any public tool.

The Practical Bottom Line

Long Island has quietly built one of the denser surveillance networks in the state, with essentially no statutory framework governing it. Nassau and Suffolk agencies ran roughly 2,800 database searches in a single month. Nassau County has announced a facial recognition and plate reader network along its border with New York City without disclosing a single operational parameter. New York requires no audits, no published retention limits, and no penalties for misuse.

For most drivers, this is an abstraction. For someone facing charges that began with a camera on a pole, it is not.

The defense posture is straightforward, and it starts with refusing to accept the police narrative that a plate reader “flagged” a vehicle. Something specific happened: a machine read characters with some confidence level, a database returned a record of some age and accuracy, and a human being made a decision. Each of those steps produces a record. New York’s discovery statute entitles you to those records, with a presumption in favor of disclosure. And the same audit trail that has exposed officer misconduct in a dozen other states will, in the right case, expose a stop that never should have happened.

If a license plate reader is the reason you were pulled over in Nassau County, get the logs. Get them early, before the retention window closes.

This article is provided for general informational purposes and does not constitute legal advice or create an attorney-client relationship. Every case turns on its own facts. If you are facing criminal charges, consult a licensed New York attorney about your specific situation. Attorney advertising. Prior results do not guarantee a similar outcome.