A software engineer drove from Washington state to Alabama and kept noticing the same object: a black pole with a small camera and a solar panel bolted to it, standing at the side of ordinary roads. He photographed one, ran the image through a search engine, and landed on the website of a company called Flock Safety.
Then he built a map.
DeFlock is a crowdsourced project that plots the locations of automatic license plate readers, and it has become the closest thing the public has to an answer to a question police departments have generally declined to answer: where are these cameras, exactly?
It has also produced a small piece of trademark law worth knowing about, because Flock tried to shut it down and lost.
What is DeFlock?
DeFlock is an open-source, volunteer-built map of ALPR camera locations, founded by Will Freeman and named, pointedly, after the company that sells most of the readers in the United States.
It runs on OpenStreetMap, the open equivalent of Google Maps, which matters for a reason that is easy to miss. The data does not sit in a private company’s database. It sits in a public one, and it can be exported, checked and reused by anybody.
Members of the public report cameras they find. Usefully, a report can record not just the position of a reader but the direction it faces, which turns a scattering of dots into something far more informative: a picture of how a department has chosen to cover a town, which approaches it watches, and which it does not.
The main map lives at maps.deflock.org.
How big is the map?
When 404 Media first wrote about the project in July 2024, it held roughly 5,600 readers worldwide, about 1,700 of them in the United States.
By December 2024 the total was approaching 12,000.
That growth is the interesting number, not the absolute figure. It doubled in months because the cameras are visible, identifiable once you know what you are looking at, and installed in public. There is no technical barrier to cataloguing them. There is only the question of whether anybody bothers.
Flock tried to shut it down
Here is the part that belongs in a legal publication.
Flock Safety sent Freeman a cease-and-desist letter. The claim was not trademark infringement. It was trademark dilution.
That distinction matters. Infringement is about consumer confusion, the idea that someone might mistake one product for another. Nobody was going to mistake a volunteer surveillance map for a surveillance company. Dilution is a different theory, available only to owners of famous marks, and it targets uses that blur or tarnish the brand.
The Electronic Frontier Foundation took Freeman’s case and refused the demand outright. Senior staff attorney Cara Gagliano wrote back: “The claims alleged in your letter are groundless, and Mr. Freeman will not be complying with your demands.” EFF has published both letters.
Why the dilution claim failed
The answer is in the statute, and it is not a close question.
Federal anti-dilution law is 15 U.S.C. § 1125(c), and subsection (c)(3) carves out three categories that are simply not actionable as dilution:
- Fair use, expressly including “identifying and parodying, criticizing, or commenting upon the famous mark owner or the goods or services”
- News reporting: “All forms of news reporting and news commentary”
- Noncommercial use: “Any noncommercial use of a mark”
DeFlock is a free, volunteer project that exists to comment critically on Flock’s product. It lands inside two of those three carve-outs at once, and arguably all three.
The broader lesson generalises well beyond one map. A company whose product is deployed in public streets does not acquire, through trademark law, the power to stop the public from describing where that product is. Naming the thing you are criticising is the ordinary mechanism of criticism.
Where are the Flock cameras in Massachusetts?
Until this month the official answer was that you were not entitled to know.
The Massachusetts State Police took the position that it could withhold both the locations of its cameras and the audit logs showing which officers had run searches. In September 2026 a Suffolk Superior Court judge rejected that, in a case brought by the ACLU of Massachusetts after five years of effort, and ordered disclosure of camera locations and search logs with plate numbers redacted. We covered that ruling and what it means separately.
So there are now two routes to the same question, and they check each other.
The crowdsourced map tells you what volunteers have physically seen. The public records process tells you what the agency admits to owning. Where those two diverge is itself information.
How to find out whether your town has ALPR cameras
Four methods, roughly in order of effort.
Check the map. Start at maps.deflock.org and look at your own roads. Absence of a dot is not evidence of absence of a camera; it means nobody has reported one.
Learn the hardware. Flock’s standard installation is a small camera on a slim pole, usually solar powered with a panel on top, mounted eight to twelve feet up and aimed along a traffic lane rather than at an intersection. Once you have identified one you will not stop seeing them.
Read the paper trail. A police department that bought cameras generated documents: a contract, a purchase order, a budget line, very often a city council or select board vote, and increasingly a written use policy. Those are public.
Make a records request. Under the Massachusetts Public Records Law, G.L. c. 66, § 10, you can ask a department directly. Useful questions: which system is in use and under what contract; how long data is retained; whether the agency has opted into national lookup so out-of-state agencies can search its cameras; whether federal sharing is enabled; how many officers hold search credentials; what reason an officer must enter to run a search; and how often the audit logs are reviewed.
That last question is the one that decides whether any of the others matter. In Revere, a police officer ran his ex-girlfriend through the city’s cameras more than a dozen times, every search was logged, and the first internal review cleared him.
What the cameras themselves capture
Worth being precise, because this is where both the alarm and the reassurance tend to overshoot.
An automatic license plate reader photographs a plate, converts the image to text, and stores a record. The Supreme Judicial Court set out the fields in 2020: a photograph of the plate, the system’s reading of the plate number, and “other data, such as the date, time, location, direction of travel, and travel lane.”
Current systems add what the industry calls a vehicle fingerprint: make, body type, colour, roof rack, bumper stickers, visible damage. That matters more than it sounds, because it means a search does not require a plate number. A description will do.
What they do not do is photograph faces or record the interior of a vehicle. This is not facial recognition, and people who claim otherwise are making the case against the technology weaker, not stronger.
The other half of the problem: bad information about Flock
A publication covering this honestly has to say the following, and most coverage does not.
The backlash to ALPRs has generated its own genre of nonsense. 404 Media, which has done more original reporting on Flock than anyone, has separately written about anti-Flock influencers who invent claims for engagement. Assertions circulate that these cameras run facial recognition, that they record audio, that they are linked to systems they are not linked to.
Every one of those makes the genuine problems easier to dismiss. And the genuine problems are not small: an officer in Revere ran a private citizen more than a dozen times and his department cleared him on first review; agencies opted into a national lookup pool without residents being told; thousands of searches were reportedly run at federal request for immigration purposes.
Those are documented. They do not need embellishment, and embellishment costs the argument its credibility with exactly the town meetings and city councils where the decisions get made.
If you are bringing this to a select board, bring the contract, the retention setting and the audit policy. Do not bring a claim you read on social media.
What Flock says
The company’s position, fairly stated, is that its cameras solve crimes, that agencies own their own data, and that the system is auditable by design. Departments do point to recovered stolen vehicles and located suspects, and there is no honest argument that the technology has no law enforcement value.
Flock has also moved substantially under pressure. It cut default retention from thirty days to seven, made search auditing compulsory, required longer holds to be tied to an active case number, added automatic keyword blocks for immigration-related and reproductive-health-related search terms in June 2025, and in January 2026 gave agencies a single toggle to switch federal sharing off entirely.
Read one way, that is a responsible vendor responding to legitimate concerns. Read another, it is a company that made those changes only after fifty jurisdictions walked. Both readings are available on the same facts, and the existence of a federal-sharing toggle does tell you what was flowing before there was one.
What the map cannot tell you
Being honest about the limits is what separates a useful tool from a talisman.
It is incomplete. It records what volunteers happened to notice. Rural coverage is thinner than urban. A blank area on the map is a statement about reporting, not about surveillance.
It can be out of date. Cameras get moved, added and removed, and a pin does not expire on its own.
It shows hardware, not access. This is the big one. Knowing where a camera stands tells you nothing about who can search what it recorded. Agencies that opt into national lookup can query each other’s data across state lines, which means the relevant question is not only which cameras are near you but which departments can see them.
And it says nothing about retention. Flock cut its default retention from thirty days to seven in 2026 after a run of documented police misuse, and made search auditing compulsory. A default is a setting. Settings are changed by whoever administers the account.
Is it legal to photograph or map a police camera?
This comes up immediately whenever somebody starts contributing, and the answer is more settled than people assume.
Photographing something in plain view from a public place is generally lawful. A camera bolted to a pole beside a road is not a secret installation, it is street furniture. Recording where public equipment sits, in public, is the same activity as noting where a fire hydrant or a speed camera is.
Publishing that information is also, ordinarily, protected expression. That is precisely the ground on which the trademark demand against DeFlock collapsed: the map is criticism and comment, and federal law expressly carves both out.
Two genuine limits are worth stating plainly.
Do not go onto private property to get the photograph. Many readers are mounted on utility poles in public rights of way, but some are not, and trespass does not become lawful because the subject is interesting.
Do not interfere with the equipment. Cutting, covering, spray-painting or knocking down a camera is not civil disobedience with a mapping project attached, it is destruction of property, and in some circumstances it is a felony. There has been a genuine wave of ALPR vandalism nationally. It is a crime, it is prosecutable, and it hands the other side of this argument its best available talking point. Nothing in this article should be read as encouraging it.
The whole strength of the mapping approach is that it is boring and lawful. That is why it worked, and why the cease-and-desist failed.
How to add a camera to the map
The contribution process is deliberately simple, because a project that depends on volunteers cannot afford friction.
Note the location as precisely as you can, and record which way the camera faces, since direction is what turns a dot into intelligence about coverage. Photograph it from the public way. Then submit it through the map interface, where the report joins the underlying OpenStreetMap data rather than a private database.
If you want your contribution to be genuinely useful to a town meeting rather than only to the map, pair it with the paperwork. A pin on a corner is an observation. A pin on a corner plus the contract, the retention setting and the audit policy is an argument.
Does mapping the cameras actually change anything?
The measurable answer is yes, and it is visible in contract cancellations.
More than two dozen Massachusetts cities and towns have opted out of the technology or stopped using it. Cambridge cancelled after cameras were already installed. Brookline paused a rollout. Watertown ended its use. Abington, Ayer, Groton and Southbridge have terminated contracts or suspended use.
Nationally, more than twenty jurisdictions moved to cancel Flock contracts in July 2026 alone, the highest total in any single month since DeFlock began tracking installations in 2021, with more than fifty cities and counties cutting ties across the year.
Those votes did not happen because residents objected to a technology in the abstract. They happened because somebody produced a specific fact about a specific street and brought it to a meeting. A map is a machine for producing that kind of fact.
The legal state of play
Massachusetts has no comprehensive statute governing automatic license plate readers. There is no general rule about who may search, what justification is required, how long data may be held, or whether it may cross state lines. What exists is each department’s own policy, the vendor’s default settings, and the public records law.
The constitutional position rests on Commonwealth v. McCarthy, 484 Mass. 493 (2020), where the Supreme Judicial Court held that four fixed readers at the Bourne and Sagamore bridges did not amount to a search, while stating clearly that a defendant has “a constitutionally protected expectation of privacy in the whole of his public movements” and that widespread ALPR use could implicate it.
Nobody has yet put a network of the current density in front of the SJC. When somebody does, the evidence of how dense that network actually is will have to come from somewhere. We have written more on what these cameras record and who can search them.
Which is the quiet significance of a volunteer map. It is not only an awareness campaign. It is a growing body of evidence about scale, built in public, in a format a court could actually look at.