No Privacy in Public: The Cases Behind Flock Cameras

Seventeen North Carolina communities have already said no to Flock. Add your name ›  ·  The Cape Fear is under warrantless AI surveillance — it’s time to act.
← All updates

“You Were in Public” Is Not the End of the Argument

The best case for these cameras is a legal one, and it has real precedent behind it. Here it is at full strength — and here is what fifteen years of Supreme Court decisions have done to it.

If you drove Market Street this morning, or crossed the Cape Fear Memorial Bridge, or turned into a shopping center off College Road, a camera recorded your plate. It logged the time, the place, and the direction you were traveling. It did that without a warrant, without suspicion, and without asking.

The best objection to our work is not about crime rates or search counts. It is a legal argument, and it is a real one:

Your license plate sits in plain view on a public road. The Supreme Court has held for sixty years that what you knowingly expose to the public is not protected by the Fourth Amendment. A camera photographing your plate therefore is not a search, and no warrant is required. Courts have upheld pole cameras watching a home for months on the same reasoning.

That argument has case law behind it. It has won, repeatedly, in courts that matter. Anyone opposing these cameras should be able to answer it honestly, or should stop talking.

Here is the answer, with the citations, including the ones that cut against us.

Steel-manning it

Where the argument is right

Katz v. United States, 389 U.S. 347 (1967), built the modern framework. Its most quoted line is the one the government relies on here: “What a person knowingly exposes to the public, even in his own home or office, is not a subject of Fourth Amendment protection.” Id. at 351. Two points of precision, because both matter. First, the “reasonable expectation of privacy” test everyone attributes to Katz comes from Justice Harlan’s concurrence, id. at 361, not the majority. Second, the sentence immediately after the one the government quotes reads: “But what he seeks to preserve as private, even in an area accessible to the public, may be constitutionally protected.” Both halves are Katz.

United States v. Knotts, 460 U.S. 276 (1983), is the government’s better case. Officers tracked a car on public roads with a radio beeper. Not a search: “A person traveling in an automobile on public thoroughfares has no reasonable expectation of privacy in his movements from one place to another.” Id. at 281.

One officer. One photograph. One public street. On those facts the plain-view argument holds, and if Flock were a single camera taking a single snapshot we would have a much weaker case. We say so plainly.

The reserved question

What Knotts left open — and what it didn’t

Knotts contained its own warning. The Court expressly declined to decide what happens when the tool scales:

“[I]f such dragnet-type law enforcement practices as respondent envisions should eventually occur, there will be time enough then to determine whether different constitutional principles may be applicable.” Knotts, 460 U.S. at 283–84.

We will be honest about the limits of that sentence, because advocates routinely oversell it. The Court did not condemn dragnets. It deferred the question. And in the very next breath it added a line our opponents should quote more often than they do: “We have never equated police efficiency with unconstitutionality, and we decline to do so now.” Id. at 284.

So Knotts is not a promise. It is an unanswered question, asked in 1983, about a thing that did not exist yet. It exists now.

Fifteen years of narrowing

What the Court has done since

United States v. Jones, 565 U.S. 400 (2012). Twenty-eight days of GPS tracking on a car. The majority, per Justice Scalia, decided it on trespass grounds — the government physically attached a device to an effect. That holding does not help us much, and we won’t pretend it does.

The concurrences are what matter. Justice Alito, joined by Justices Ginsburg, Breyer, and Kagan, wrote that “the use of longer term GPS monitoring in investigations of most offenses impinges on expectations of privacy.” Id. at 430. Justice Sotomayor, concurring separately, went further: GPS monitoring “generates a precise, comprehensive record of a person’s public movements that reflects a wealth of detail about her familial, political, professional, religious, and sexual associations.” Id. at 415.

Be careful how you count that. Those are two separate writings, one concurring only in the judgment — there is no five-Justice holding on aggregation in Jones. But the Court itself later did the arithmetic: “A majority of this Court has already recognized that individuals have a reasonable expectation of privacy in the whole of their physical movements.” Carpenter v. United States, 585 U.S. 296, 310 (2018).

Carpenter v. United States, 585 U.S. 296 (2018). The government obtained historical cell-site location records. A search — even though every individual data point had been “shared” with the phone company. Chief Justice Roberts: cell phone location information is “detailed, encyclopedic, and effortlessly compiled.” Id. at 309.

Now the part most advocacy pieces skip. Carpenter is expressly narrow, and it contains a carve-out written for the other side:

“Our decision today is a narrow one. We do not express a view on matters not before us… We do not disturb the application of Smith and Miller or call into question conventional surveillance techniques and tools, such as security cameras.” Id. at 316.

That is the sentence every court that has rejected an ALPR challenge has reached for. It is a fair reading of the opinion. Our answer is not that the carve-out doesn’t exist — it does — but that a networked, searchable, retroactive plate database is not what anyone in 2018 meant by “a security camera,” any more than a phone company’s tower records were what anyone in 1979 meant by a pen register. Whether courts accept that is not yet settled.

One more correction while we’re being precise: Carpenter did not announce a seven-day rule. It held that “accessing seven days of CSLI constitutes a Fourth Amendment search” while expressly declining to decide “whether there is a limited period for which the Government may obtain an individual’s historical CSLI free from Fourth Amendment scrutiny, and if so, how long that period might be.” Id. at 310 n.3.

The newest case

Chatrie, stated accurately

Chatrie v. United States, No. 25–112, 609 U.S. ___ (June 29, 2026), is three weeks old, and it is already being misdescribed — including, in an earlier draft, by us. So here is what it actually says.

What it held. Justice Kagan, for the Court: police “conducted a Fourth Amendment search when they acquired Chatrie’s location data from Google because an individual has a reasonable expectation of privacy in his cell-phone location information.” On the third-party doctrine: “just as the third-party doctrine did not apply in Carpenter, it does not apply here,” because a phone user “is not to be viewed as sharing private information with third parties… just by doing the ordinary things cell-phone users do.”

What it did not hold. Chatrie did not rule that geofence queries require a warrant backed by probable cause. It decided only that acquiring the data was a search, and sent the reasonableness question back: “The conclusion that a Fourth Amendment search occurred does not resolve this case, because the Fourth Amendment prohibits only searches that are ‘unreasonable.’” Whether the warrant was valid at each step goes to the Fourth Circuit in the first instance. The Court also expressly reserved the good-faith question — meaning Okello Chatrie may still lose.

The vote. Six to three on the judgment; the opinion of the Court commanded five. Justice Gorsuch concurred in the judgment only, reaching a search by a property route and discarding Katz altogether. Justice Alito dissented, joined in part by Justices Thomas and Barrett; Justice Barrett dissented separately.

And it says nothing about license plates. Search the entire opinion — majority, concurrences, dissents — for “license plate,” “plate reader,” or “ALPR” and you will find zero matches. Anyone telling you Chatrie decided anything about Flock is selling something.

What Chatrie gives us is reasoning, not authority. And one piece of that reasoning matters here more than the rest: the Court rejected the argument that pulling a narrow time-slice out of an enormous database escapes the Fourth Amendment. Protection does not switch on only once an intrusion “goes too far,” and that principle, the Court said, “makes all the more sense when… law enforcement officials can select the time-limited set of materials they want from an all-encompassing database.” Read that sentence and then read a Flock search log.

The actual difference

Why a network is not an officer on a corner

This is the part the doctrine turns on, and it is worth saying without any law at all.

A human officer watching Market Street has limited memory, limited endurance, and can be in one place. He forgets. He goes home. He cannot tell you where your car was on a Tuesday in April, because he wasn’t looking, and if he had been, he would not remember.

A network has perfect memory, no endurance limit, and is everywhere at once. It does not forget. It never goes home. It can tell you where your car was on a Tuesday in April in about four seconds, along with every other Tuesday, and it can do that for you and for everyone else simultaneously, because nobody had to be assigned to watch you.

That is not the same tool used more efficiently. Scale is the constitutional argument, and it is the argument Jones, Carpenter, and Chatrie have each, in their own way, taken seriously. Scholars call it the mosaic theory: thousands of individually innocent observations can, assembled, reveal what no single observation could. One tile is not a picture.

An officer seeing your car on Market Street learns nothing. A database that records your car on Market Street every weekday at 7:40, in the church lot on Sunday, outside a clinic on the 14th, and on a particular block last Tuesday has assembled something no officer on a corner ever could.

Read this part

The best arguments against our position

If you only read one section, read this one. These are the authorities we would have to beat, stated as strongly as we can state them.

1. The pole-camera cases, and courts keep winning them. In United States v. Tuggle, 4 F.4th 505 (7th Cir. 2021), officers watched a home with three pole cameras for roughly eighteen months without a warrant. Not a search. The Seventh Circuit considered the mosaic theory and rejected it. In United States v. Bucci, 582 F.3d 108 (1st Cir. 2009), eight months of pole-camera surveillance of a driveway: not a search. In United States v. Moore-Bush, 36 F.4th 320 (1st Cir. 2022), the First Circuit took eight months of camera surveillance en banc — and the defendants lost 6–0 on the judgment, with the court splitting 3–3 on the reasoning and Bucci left standing. These are not close calls we can wave away. They are the mainstream.

2. Carpenter carved out security cameras. Discussed above. It is in the opinion, in the narrowing paragraph, in the Chief Justice’s own words. Every court rejecting an ALPR claim cites it.

3. The third-party doctrine still exists. Smith v. Maryland, 442 U.S. 735 (1979), holds there is no reasonable expectation of privacy in information voluntarily conveyed to a third party. Carpenter and Chatrie declined to extend it; neither overruled it.

4. Plate checks are old. Officers have run plates against NCIC for decades and nobody thinks that needs a warrant. The government’s framing is that ALPRs simply automate a lawful practice.

5. Public observation has long been permitted. California v. Ciraolo, 476 U.S. 207 (1986) (naked-eye aerial observation of a fenced yard from 1,000 feet: not a search); Florida v. Riley, 488 U.S. 445 (1989) (plurality) (helicopter at 400 feet: not a search). Courts have tolerated a great deal of watching from public vantage points.

6. And on ALPRs specifically, we are losing. This is the one to sit with. No appellate court has ever held that a networked ALPR system requires a warrant. Not one. Massachusetts came closest in Commonwealth v. McCarthy, 484 Mass. 493 (2020), which adopted an aggregation principle in reasoning — and then ruled against the defendant on four cameras at two bridges. A Norfolk trial judge did require a warrant in Bell v. Commonwealth, 113 Va. Cir. 316 (2024), and that ruling is effectively dead: never appealed, and repudiated when the Virginia Court of Appeals reversed the same judge in a later case.

What actually binds us

North Carolina sits in the Fourth Circuit. Here is what that means.

Most of what you just read is persuasive authority here, not binding. So what binds North Carolina?

On pole cameras: nothing. The Fourth Circuit has not squarely decided whether long-term warrantless pole-camera surveillance is a search. The nearest case, United States v. Vankesteren, 553 F.3d 286 (4th Cir. 2009), involved a camera in open fields — a different doctrine — and predates both Jones and Carpenter. As Jeff Welty of the UNC School of Government put it in 2022: “I don’t think that the Fourth Circuit has a post-Carpenter case on point.” No North Carolina appellate court has decided one either. On this question, our circuit is a blank page.

On mass surveillance: one case, and it is ours. In Leaders of a Beautiful Struggle v. Baltimore Police Department, 2 F.4th 330 (4th Cir. 2021) (en banc), the Fourth Circuit held that a warrant was required for an aerial surveillance program that covered roughly 90% of the city. It is not an ALPR case, and courts elsewhere routinely distinguish it. But it is the one binding decision in this circuit holding that coverage of a whole city crosses a constitutional line, and it is the best authority a North Carolina driver has.

And a second, from September 2026. In Andrews v. Baltimore City Police Department, No. 18-1953 (4th Cir. Sept. 24, 2026), the Fourth Circuit held that using a cell-site simulator to locate a person is a Fourth Amendment search. It is not an ALPR case either — the officers even won on qualified immunity — but it is the circuit saying again that a technology which lets the government reconstruct where someone is, in a way ordinary observation cannot, is a search. That is the exact point the “you were on a public road” defense has to overcome: one camera sees you in public; a searchable, networked record of every camera reconstructs your movements, and that is different in kind.

And the case to watch is being briefed right now. In Schmidt v. City of Norfolk, two Norfolk residents challenged a network of 176 Flock cameras in 75 clusters. We should be accurate about where it stands, because the earlier version of this post was not. Chief Judge Mark Davis denied the City’s motion to dismiss in February 2025 — but on January 27, 2026, he granted summary judgment for Norfolk. The plaintiffs lost. His reasoning: “Plaintiffs are unable to demonstrate that defendants’ ALPR system is capable of tracking the whole of a person’s movements.”

Read the rest of that opinion, though, because it is not a celebration of the technology. Judge Davis also wrote that the database “is not limited to individuals engaged in criminal activity based on probable cause, or even reasonable suspicion, but instead ensnares an immense array of law-abiding citizens guilty of nothing more than driving their cars in Norfolk.” And on whether ALPR surveillance could ever become too intrusive: “While a definitive answer to that question is elusive, what is readily apparent to this Court is that, at least in Norfolk, Virginia, the answer is: not today.”

Not today is doing a great deal of work in that sentence. It is an invitation to a different record — more cameras, longer retention, broader sharing.

The case is now on appeal to the Fourth Circuit as No. 26–1227, fully briefed, not yet argued. The plaintiffs are represented by the Institute for Justice, not the ACLU, though the ACLU, EFF, EPIC, and Cato have filed as amici on their side. And on Norfolk’s side, among the amici, sit the North Carolina Association of Chiefs of Police and the North Carolina Sheriffs’ Association.

Sit with that last one. North Carolina’s law enforcement associations have gone to the Fourth Circuit — the court that binds New Hanover County — to argue that scanning drivers who are suspected of nothing is not a Fourth Amendment search. They understand exactly what is at stake in that courtroom. So should you.

The argument nobody has made

North Carolina’s constitution uses different words

Everything above is federal law. But New Hanover County is not only bound by the Fourth Amendment, and North Carolina’s search provision is not a copy of it. Read it:

What follows is the short version. The long one, written for licensed counsel and stating the government’s side as fairly as ours, is our Memorandum of Law on the North Carolina Constitution and warrantless ALPR/PTZ surveillance (PDF).

“General warrants, whereby any officer or other person may be commanded to search suspected places without evidence of the act committed, or to seize any person or persons not named, whose offense is not particularly described and supported by evidence, are dangerous to liberty and shall not be granted.” N.C. Const. art. I, § 20.

Notice what is not in there. No “unreasonable searches and seizures.” No “probable cause.” Those phrases appear nowhere in § 20. It is a general-warrants clause and nothing else — carried almost verbatim from § XI of North Carolina’s 1776 Declaration of Rights, written by people with the writs of assistance fresh in memory. When it was brought forward, the drafters made one change worth noticing: “ought not to be granted” became “shall not be granted.”

Now hold the clause against the thing itself. A network that scans every passing car searches without evidence of any act committed. It captures people not named. Their offense is not particularly described, because there is no offense. It is not supported by evidence, because there is no evidence — that is the design. The system collects first and finds the suspect later.

That is not an analogy. That is the clause read against the deployment, word by word. And North Carolina’s Supreme Court has itself said the provision’s language “differs markedly from the language of the Fourth Amendment.” State v. McClendon, 350 N.C. 630, 635, 517 S.E.2d 128, 132 (1999) (quoting State v. Arrington, 311 N.C. 633, 643, 319 S.E.2d 254 (1984)). The Court has the power to read our constitution independently; the federal Constitution is a floor, not a ceiling. Virmani v. Presbyterian Health Servs. Corp., 350 N.C. 449, 475, 515 S.E.2d 675, 692 (1999).

The hard part

And North Carolina’s courts have shown no appetite for it

We would be doing exactly what we accuse others of if we stopped there. So here is the other side of the state argument, and it is worse than the federal one in some respects.

State v. Garner, 331 N.C. 491, 417 S.E.2d 502 (1992) is directly against us: “there is nothing to indicate anywhere in the text of Article I, Section 20 any enlargement or expansion of rights beyond those afforded in the Fourth Amendment.” Id. at 506, 417 S.E.2d at 510. The same passage warns that § 20 “should not be viewed as a vehicle for any inventive expansion of our law.”

State v. Rogers, 920 S.E.2d 775 (N.C. Oct. 17, 2025) is more recent and more discouraging. Nine months ago the Supreme Court of North Carolina overruled State v. Carter, 322 N.C. 709, 370 S.E.2d 553 (1988) — the one case that had ever read § 20 to give North Carolinians more than the Fourth Amendment gives. Chief Justice Newby called Carter a “confused opinion” whose reasoning was “dubious.” The Court assumed without deciding that § 20 carries any exclusionary rule at all — and then held that whatever rule it might carry contains a federal-equivalent good-faith exception. The vote was 5–2, Justice Earls dissenting with Justice Riggs. And the case came out of New Hanover County.

So state the score honestly: as of today there is no surviving North Carolina decision construing Article I, § 20 more protectively than the Fourth Amendment. Carter was the only one. It is gone. And when a § 20 general-warrant challenge was last aimed at a surveillance program — satellite-based monitoring, in State v. Hilton, 2021-NCSC-115 — it lost.

The answer

Read the rest of Garner‘s sentence

Here is the whole sentence from Garner, including the parenthetical that almost every quotation drops:

“[I]t is abundantly clear that the language of this provision of our Constitution, relating entirely to ‘general warrants,’ of the past (while still relevant to protect against any recurrence of the historic abuses specified), should not be viewed as a vehicle for any inventive expansion of our law.” Garner, 331 N.C. at 506, 417 S.E.2d at 510.

Garner did not read § 20 out of the constitution. It reserved § 20’s force against any recurrence of the historic abuses specified — and then told courts not to stretch it past that.

So what is the historic abuse specified? The clause says it: authority to search without evidence of the act committed, sweeping up persons not named whose offense is not particularly described. Asking whether a suspicionless dragnet is that thing is not an inventive expansion of our law. It is the only job Garner left the clause, and no North Carolina court has ever been asked to do it here.

And the ALPR statute makes the point better than we can. Article 3D of Chapter 20 does require particularized facts — a “written, sworn statement” setting out “specific and articulable facts showing that there are reasonable grounds to believe” the data matters. N.C.G.S. § 20-183.32(c). But read where that requirement attaches: only to preserving data past ninety days. At capture, nothing. No provision in Article 3D limits which plates may be scanned. At routine query, a written request naming a purpose category — and “[w]ritten requests may be in electronic format.” § 20-183.32(e). No facts about any person are required at either point.

The General Assembly knew exactly how to demand particularized, sworn facts. It wrote the standard into the same article. It simply did not attach it to the moment your car is scanned, or to the moment someone looks you up. That gap is the whole argument, and it is the county’s to explain.

Where that leaves the state claim. Article I, § 20 has never been applied to an ALPR network by any North Carolina court, published or unpublished. It is a blank slate. That is not the same as a strong hand — a blank slate cuts both ways, and Garner and Rogers tell you which way this Court is currently leaning. But it is the one argument in this post that has not already been tried and lost, and it rests on words that are actually in our constitution rather than on an analogy to somebody else’s phone records. The vehicle would not be a suppression motion, which Rogers now makes a poor bet, but a civil action for declaratory and injunctive relief — where sovereign immunity does not bar a colorable direct claim under the state constitution. Corum v. Univ. of N.C., 330 N.C. 761, 782, 413 S.E.2d 276, 289 (1992); Kinsley v. Ace Speedway Racing, Ltd., 904 S.E.2d 720 (N.C. 2024).

That road has a gate of its own, and we should name it rather than let you find it. A plaintiff who already holds remedies that are “meaningful, even if not necessarily complete or the relief they want,” cannot bring a direct constitutional claim at all. Washington v. Cline, 385 N.C. 824, 830, 898 S.E.2d 667 (2024). Our answer is that a Class 1 misdemeanor and after-the-fact damages neither stop ongoing collection nor redress a forward-looking, systemic injury — but that is an argument, not a holding, and the county would contest it. What is settled is narrower: exhausting administrative remedies is not a jurisdictional prerequisite, so the county could not get such a case dismissed at the threshold. Askew v. City of Kinston, 902 S.E.2d 722 (N.C. 2024); accord Singleton v. N.C. Dep’t of Health & Human Servs., 906 S.E.2d 806 (N.C. 2024). We work the whole vehicle question through — counts, standing, relief, and every obstacle above — in our memorandum of law for North Carolina counsel.

The eight nobody asks about

The Condors are a different question entirely

One more thing, and it may be the most useful paragraph here.

Every argument above — ours and the government’s — is about plate readers. Fixed cameras, automated, photographing what passes. That is what Schmidt was about. That is what Tuggle and Bucci were about. That is what Carpenter‘s “security cameras” carve-out contemplates.

New Hanover County’s contract does not only buy plate readers. It buys 8 Condor live-video cameras — pan-tilt-zoom units that a human being points — alongside the plate readers. By the Sheriff’s own August 2026 count the fleet is 36 cameras: 28 plate readers + 8 Condor.

A Falcon is automated. It photographs whatever drives by, and no one chose you. A Condor is aimed. It is live video, it swivels, it zooms, and a person decides where it looks and for how long. That is not plate reading. Whatever else it is, it is not the “conventional surveillance tool” the Chief Justice had in mind, and it is not a camera bolted to one spot staring at one driveway.

Which means the Condors do not need the mosaic theory at all. A discretionary, human-aimed, zooming video camera raises ordinary Fourth Amendment search questions — and if one can see a home, curtilage questions that are a century old and thoroughly settled. None of the ALPR cases reach it, because none of them were about this.

Nobody has asked the county what the eight Condors point at, how often they are moved, who authorizes a move, whether any is logged, or whether any of them can see a house. Those questions have answers, the answers are in records the county holds, and not one of them requires a court to extend anything.

Honest bottom line

What we can say, and what we can’t

We cannot tell you that Flock cameras are unconstitutional. No appellate court has held that networked ALPRs require a warrant, the pole-camera cases run against the mosaic theory, Carpenter expressly preserved security cameras, and the only judge to rule on a network this size ruled for the city. Anyone who tells you the law is settled in our favor is not reading the same cases we are.

We should also be precise about the technology, since we ask the county to be. Flock deployments are not identical. Retention periods, sharing settings, and search permissions vary between agencies, and an argument that describes the broadest possible configuration as universal deserves to lose. What we know about this county’s configuration — at least 62 mapped fixed cameras, 8 aimable Condors, a 90-day cap, and 2,980,082 searches in sixteen months — we know because we asked for the records and read them. Whether that record is denser than Norfolk’s is a factual question, and it is the question the next case turns on.

Here is what we can say. The public-exposure doctrine has narrowed for fifteen consecutive years, every single time the government has argued that technology let it assemble what an officer could once only glimpse. Jones in 2012. Carpenter in 2018. Chatrie three weeks ago. The Supreme Court has grown steadily more concerned about aggregated location surveillance. But lower courts have not extended that concern to ordinary ALPR networks, and on the cases decided so far, they are declining to.

These networks were built on the legal assumptions of 1983. They are being tested under the law of 2026. Knotts promised that when the dragnet arrived, there would be “time enough then” to decide whether different principles apply.

It arrived. It is on Market Street. That question is now in front of the Fourth Circuit, and the only people who have shown up in that courtroom to speak for North Carolina drivers are not from North Carolina.

You do not have to wait for a court. A county commission can decline to pay for it on any Tuesday it chooses — and one just did, five to nothing, five hours west of here.

How to check us. Every case above is cited so you can read it yourself, including the six that cut against us. If we have a holding wrong, we want to know — write to Mark@DeFlockILM.org and we will correct it on this page and say what changed. Chatrie pin cites are to the slip opinion; the U.S. Reports pagination is not yet available. Our fuller analysis for practitioners is in two memoranda — the North Carolina Constitution memo (PDF) and the § 1983 memo (PDF) — alongside the rest of the attorney resources. It is a working document and we want it marked up.

Disclosure. DeFlockILM is a group of Wilmington citizens who want our government to respect the U.S. and North Carolina Constitutions and stop mass surveillance. This post is public legal information, not legal advice, and no court has adopted the argument we make here.

Your move

You’re not a suspect. So stop being tracked like one.

It takes one minute. Add your name — then tell the officials who represent you to take the cameras down.