Security & Surveillance

Do ALPR Searches Need a Warrant? Sturdivant and Footage Search

EFF and the ACLU asked the Sixth Circuit to require a warrant for ALPR searches. Here is the case, the split among courts, and what it means for footage review.

Ceptory Team8 min read
A search bar above four video results with the matching night perimeter clip highlighted

TL;DR

On November 9, 2026, EFF and the ACLU filed an amicus brief in the Sixth Circuit arguing that searching ALPR databases is a Fourth Amendment search that needs a warrant. Courts are split, so teams that search footage should log every search and its purpose now.

On November 9, 2026, the Electronic Frontier Foundation and the ACLU filed an amicus brief in the Sixth Circuit asking the court to hold that searching automated license plate reader (ALPR) databases is a Fourth Amendment "search" that requires a warrant. The case is United States v. Sturdivant. For anyone who reviews surveillance footage, the question behind it is practical: when does searching a large body of recorded video stop being routine review and become something a court wants a warrant for? The answer today is that courts disagree, and the way a search is run is part of the argument.

What did EFF and the ACLU file in Sturdivant?

They filed an amicus brief arguing that police tracking of a suspect through ALPR data was a Fourth Amendment search. According to EFF's announcement, the case "involves a series of robberies in and around Cleveland, Ohio." Police used surveillance camera footage to identify the robber's car by make, model and color, but did not have the plate number. They then "ran an open-ended search of two ALPR databases, Flock and ELSAG," and learned that a car belonging to the defendant had been tracked near the robbery sites, which led to his arrest.

The district court denied suppression. It held that Mr. Sturdivant had no reasonable expectation of privacy in a license plate number visible to the public, and so no Fourth Amendment interest in not having his car tracked in public places. EFF disagrees, writing that ALPRs give police "a virtual time machine to view suspects', or anyone else's, past movements (and much more) with just the click of a button." The brief asks the Sixth Circuit to reverse and hold that ALPR searches require a warrant.

Note the shape of the investigation: footage first, database second. A description pulled from video (make, model, color) became the key for a search across a network. That chain is exactly how modern investigations work, and it is why the case matters beyond license plates.

Are courts agreeing on ALPR and the Fourth Amendment?

No. Recent decisions point in opposite directions.

DecisionCourtDateOutcome
Schmidt v. City of NorfolkE.D. Va. (Judge Mark Davis)Early 2026Upheld Norfolk's roughly 200 Flock cameras; warned the balance "could conceivably tip the other way" as networks grow
United States v. KyleN.D. Okla. (Judge Sara Hill)Oct. 1, 2026Held that using an extensive ALPR system required a warrant
United States v. Sturdivant6th Cir.Amicus filed Nov. 9, 2026Pending

The Norfolk ruling, reported by Carscoops, rested on case law that drivers have no reasonable expectation of privacy in plates displayed on public roads, and said courts have "almost uniformly concluded" that photographing plates and storing them in a searchable database is not a warrantless search. The judge cautioned that as retention periods grow and analytic tools improve, the balance could change. The plaintiffs planned to appeal.

Lawfare's analysis describes Kyle as the first federal ruling declaring warrantless ALPR use a Fourth Amendment violation. A Tulsa deputy ran an SUV's plate and got more than 50 sightings across multiple states over 30 days; the court treated that as a detailed log of the vehicle's movements, relying on Kyllo and Carpenter. Lawfare also notes earlier decisions went the other way, including a Massachusetts high court ruling from 2020. EFF's post says a federal district court recently called ALPRs "a form of indiscriminate mass surveillance" that can violate the Fourth Amendment.

Why does this matter for people who search CCTV footage?

Because the legal reasoning keys on aggregation and retrospective search, not on any one camera. The common thread in the cases above is that a single sighting of a plate is unremarkable, but a searchable, long-retained record that lets someone reconstruct movements is different. The Norfolk judge's warning is about retention and analytic tools. The Oklahoma judge focused on 30 days of sightings.

Most corporate and municipal CCTV is not an ALPR network, and these cases concern police. Nothing here says a warehouse operator needs a warrant to review its own camera. But three practical lessons carry over:

  1. Scope and duration are part of the story. A narrow search of one site for one incident window reads very differently from an open-ended search across months and locations.
  2. Retention is a legal variable. The longer footage and derived data are kept, the more a court may treat the archive as a tracking tool. Our guide on how much storage CCTV cameras need shows how retention drives cost, and the same dial drives exposure.
  3. Documented purpose protects everyone. If you can show who searched, for what incident, with what query, you can defend the search and reproduce it.

How do investigations move from footage to database searches?

Typically in three steps: describe, filter, confirm. Investigators watch or search video for a descriptor, narrow the time window, and only then query other systems. Our guide to searching CCTV footage for a person, vehicle or event walks through the first two steps, and finding a moment in hours of CCTV footage covers narrowing the window.

Sturdivant is a vivid example of the third step going wrong, at least in the defense's view: the descriptor (a car's make, model and color) was legitimately observed on camera, but it was then used as a key across two commercial databases without a warrant. The legal fight is about that second hop, not the footage review.

What is the difference between reviewing footage and searching a network?

Reviewing footage is bounded by the camera, the site and the incident; searching a network is bounded only by the query. A camera at a loading dock records a fixed field of view, usually kept for a set number of days, and the person reviewing it already knows where and roughly when something happened. An ALPR network, in contrast, records every passing vehicle across many locations and lets a user ask where a given car has been over weeks. That is the aggregation concern in EFF's brief, and it is why the Oklahoma court cared that one plate produced more than 50 sightings in multiple states over 30 days.

For a business, the lesson is to keep the two modes separate in policy. Incident review means a named incident, a defined camera set and a short window. Open-ended pattern searching, such as "show every time this vehicle has visited any of our sites this year", is a different activity with different risks, and may be restricted by privacy law in some regions. Decide in advance which of the two your tools are used for, and who may do the second.

A simple test helps: could you explain, in one sentence, why this specific search was necessary and why it was limited to this scope? If the answer is "we wanted to see what came up", the search is probably too broad.

How can teams review footage defensibly while the law is unsettled?

Treat every search as something you may have to explain later. A defensible workflow does not need exotic tooling:

  1. Define the incident and the window before searching: date, location, camera set.
  2. Write the query down. A natural-language question such as "white pickup truck entering the north gate" is self-documenting in a way a manual scrub is not.
  3. Keep an audit trail of who ran which search and when. Ask any vendor, ours included, what it logs before relying on it.
  4. Export evidence, not the whole archive. Pull the clips that matter and record their source recorder and time range. The incident report template and our guide on writing an incident report from video footage give a structure for this.
  5. Set retention deliberately. Use the CCTV storage calculator to see what each extra week costs, then decide whether you need it.
  6. Ask counsel how your state treats ALPR, biometric and footage searches. Several states have their own statutes and constitutions that go further than federal law.

Where does Ceptory fit, and where does it not?

Ceptory is a video intelligence platform: you upload recorded video or add it by URL through the API, and search it in natural language across scenes, speech and on-screen text. It is built for reviewing specific footage, not for operating a plate-reading network, and it does not connect to live camera streams or sell cameras. That makes it a reasonable fit for the "describe and narrow" step of an investigation, run against clips you are entitled to hold. It is not a substitute for legal advice, and it does not decide whether a search is lawful. If you are weighing it against camera-platform search features, the government and security solution page and the Verkada comparison lay out the differences. Pricing is on the pricing page.

If your interest is the privacy of people on camera rather than search policy, redaction of faces and plates is a separate task; BGBlur covers it, including its own take on the ALPR warrant requirement.

What should you do now?

  1. Read EFF's post and the Lawfare summary so you know the current split.
  2. Inventory every system your team searches (cameras, ALPR vendors, shared databases) and who has access.
  3. Write a one-page search policy: allowed purposes, who may search, how results are logged.
  4. Shorten retention where there is no business reason to keep footage beyond a set window.
  5. Watch the Sixth Circuit: a ruling that ALPR searches require a warrant would likely push agencies to document purpose and scope for every search.

The bottom line

Sturdivant will not settle the law by itself, and a circuit ruling would bind only the Sixth Circuit's states. But the direction of the argument is clear: courts are paying attention to how footage and derived data are searched, how long they are kept, and whether a descriptor from video becomes a key into a much larger system. Teams that search footage can get ahead of that by being narrow, documented and short on retention. To see natural-language search on your own recorded clips, start at Ceptory.

Frequently asked questions

Do police need a warrant to search ALPR data?

It is unsettled. EFF and the ACLU argue in the Sixth Circuit that they do. A federal court in Oklahoma agreed on October 1, 2026, while a Virginia federal court upheld Norfolk's cameras in January 2026.

What is United States v. Sturdivant about?

It is a Cleveland-area robbery case. Police identified the robber's car from surveillance footage, then ran an open-ended search of the Flock and ELSAG databases and found a car belonging to the defendant near the robbery sites.

Does this ruling affect private CCTV footage I review?

Not directly. The case concerns police searches of ALPR networks. But it shows courts weighing how footage is searched and kept, so documenting purpose, scope and retention is sensible for any team.

Does Ceptory read license plates?

Ceptory searches recorded video you upload or add by URL, including text visible on screen. It is not an ALPR network and does not build a cross-site vehicle database.

What should security teams do while the law is unsettled?

Write a search policy, log who searched for what and why, set short retention, and ask counsel how your state treats ALPR and footage searches before relying on results.

Ceptory

Search your own footage in plain language. The free tier indexes 60 minutes of video.