Flock Safety and ALPR Cameras: Master Index

Who Owns the Cameras, and Who Can Search Them

This is the most common question about Flock and the easiest to answer wrongly, because three different parties are involved and the word "owns" means something different for each of them.

The vendor is Flock Safety, an American video surveillance company headquartered in Atlanta and founded in February 2018^[4]. It is a company, not a government agency and not a public utility, which is why the cameras are described throughout this index as private infrastructure serving public bodies.

The subscriber is whoever pays: a police department, a city, a homeowners association, a business. The subscriber chooses where cameras go and who inside its own organisation may run a search. This is why a resident who objects to a camera has to petition a council rather than the company, and why a cancelled contract takes cameras off a street rather than transferring them to somebody else. The free trial playbook traces how that arrangement usually begins.

The searcher is the third party, and the one that generates most of the coverage indexed below. A record captured by one subscriber's camera has repeatedly proved reachable by agencies elsewhere, including federal ones. How the national lookup works covers the design, and the immigration screening toggle covers the sharpest example of it. So if the question behind "who owns the cameras" is really "who can see where my car has been", the answer is set by sharing settings rather than by whose name is on the pole.

How to Read This Cluster

ALPR coverage is unusually easy to get wrong, because three different kinds of numbers circulate and none of them measure the same thing. Keeping them apart is most of the work.

Vendor deployment figures describe how many cameras a company says it has installed. Agency records, usually obtained through public-records requests, describe what one department's cameras did: how many scans, how many searches, who ran them. Crowdsourced map counts describe how many camera locations volunteers have found and published. A deployment figure and a map count will never reconcile, because one counts hardware a company sold and the other counts points a person walked past and logged.

A worked example, since this site got it wrong until today. OpenStreetMap carried 138,351 features tagged with a surveillance type of ALPR worldwide as of 9 August 2026, and 127,878 of those sat in the United States extract^[1][2]. That is a crowdsourced count, spanning every vendor, not Flock alone. A figure of 336,000 is sometimes attached to that dataset and does not belong to it: it is a Flock deployment estimate, weighted very differently by country. A page on this site carried that mix-up until it was corrected on 9 August 2026. When a number on any page here does not name what it counts, treat it as unverified.

The second thing worth holding onto is the base rate. A review of agency data by the Electronic Frontier Foundation found that about 0.5% of plate scans were attached to a hot list, across roughly 2.5 billion scans reported by 173 agencies^[3]. Almost everything these cameras record is a record of somebody who is not suspected of anything. Most of the policy fights indexed below are arguments about what may be done with that remainder.

Finally, watch the dates and the counting method. Cancellation tallies disagree because some count contracts and some count cities, and a city can cancel, partially restore, and cancel again. A 2025 article and a 2026 news item about the same place can both be accurate. Where two pages here differ, the later date is the better guide to the current state, and the earlier one is the better guide to how it got there.

Start Here

1. What Flock Is and How the Network Works

The foundations. Read anything in this group before reading a specific city fight, because the mechanics repeat everywhere.

• Flock Safety: The Surveillance Network Tracking Your Car: the company profile, its valuation, and its expansion from plate cameras into drones and microphones.
• How the National Database Actually Works: the shared lookup that lets one town's camera answer another state's query, and the funding behind it.
• Scan Volume and the Shared Lookup: what monthly scan totals mean, and how agencies reach each other's data.
• License Plate Readers: The Network Tracking Every Car: where plate data sits inside the wider vehicle-tracking picture, including private databases.
• The Free Trial Playbook: the sales pattern that puts cameras in a town before any council votes on them.
• FreeForm AI Searches for People, Not Just Cars: the feature that extended plate search into descriptions of human beings.
• Border Patrol's Nationwide Plate Dragnet: the federal reader network that operates well away from any border.

2. Finding and Identifying Cameras Near You

This site publishes guides on recognising the hardware and using someone else's map. It does not publish camera locations of its own.

• What Is DeFlock? The Crowdsourced ALPR Camera Map: what the project is, why deflock.me redirects to deflock.org, and what the map cannot tell you.
• How to Find and Avoid Flock ALPR Cameras Near You: what a unit looks like on a pole, why plate covers are illegal and counterproductive, and which civic levers work.
• How One Mapper Sparked a Revolt: the origin of the crowdsourced map and the legal threat it survived.
• Readers Disguised as Trailers and Barrels: the deployments built specifically not to look like cameras.

3. Who Gets the Data: ICE and Federal Access

Federal access is the single issue behind most of the cancellations in section 5. These pages document the mechanism rather than the reaction to it.

• The License Plate Readers Helping ICE Track Immigrants: the original account of immigration searches running against local cameras despite stated policy.
• Local Cameras as ICE Screening Tools: the administrative toggle that matches every passing plate against a federal immigration category.
• Class Action Over Federal Searches of California Plate Data: the suit alleging out-of-state and federal agencies searched one city's cameras at scale.
• The Doorbell to Deportation Pipeline: how consumer doorbell footage was to be joined to the plate network.
• School Cameras Feeding Immigration Enforcement: audit logs showing how often school-adjacent cameras were searched, and by whom.

4. Error, Misuse, and Security Failure

Separate from the policy question of what the system is for, there is the operational question of whether it works and who can abuse it. These are the documented failures.

• Error Rates and Interstate Tracking: how often plates are misread, what a false hit does to an innocent driver, and the abortion-tracking case.
• Police Chiefs Using the System to Stalk: repeat personal lookups, the pattern behind them, and why warrants are the proposed fix.
• Body-Cam Footage of a Deputy Stalking a Woman: a single case with unusually complete documentation of how the tools were used.
• The Same Abuse Pattern in Two Countries: British and American cases that suggest the failure mode is structural, not local.
• Cameras Streaming Online Without a Password: the security research that found live feeds open to anyone, and the congressional response.

5. What Communities Have Actually Done

This is the part of the record with the clearest cause and effect. Nearly every reversal started with a records request and a council meeting, and the losses are as instructive as the wins.

• 23 Communities Told the Cameras to Get Out: the 2025 baseline, and the organising that produced it.
• Eight Cities That Beat Flock and ICE: the playbook, broken into the steps that recur across all of them.
• The Running List of Cancellations: which cities have ended contracts and what triggered each one.
• Cities Pulling the Plug Over Federal Access: the unauthorised federal searches that turned procurement decisions into political ones.
• Silicon Valley Cancels: what Mountain View found in its own access logs before terminating.
• Oregon Cities Reject Flock: several cancellations in one state, and the statewide rules they prompted.
• Seattle Pauses and Shuts Cameras Off: an executive-level pause rather than a council vote, and what it covered.
• Berkeley's Expansion Vote: what was on the table, including drones and investigative software.
• Berkeley Delays After the Mayor Reverses: what all-night public testimony changed, and what it did not.
• Approved Despite Hours of Opposition: a loss, and the procedural move that produced it.
• Two Drone Programs, No Council Sign-Off: what happens when acquisition skips the oversight step entirely.
• A New Oversight Law, One Month In: the first stress test of a surveillance-oversight ordinance after a violent incident.
• The Ring Partnership Killed by Public Outrage: the clearest case of consumer pressure cancelling an integration before launch.
• Cameras Destroyed Across Five States: reporting on physical destruction as a backlash indicator. This site documents it and does not recommend it; the lawful routes are set out in the guides above.

6. The Legal and Policy State

Litigation and legislation are moving faster than either the vendor or the agencies. Nothing below is legal advice, and state law changes without notice.

• The San Jose Mass Surveillance Suit: the civil-liberties challenge to warrantless searches of a citywide camera network.
• A Federal Class Action Aimed Higher: the parallel case framed as a Fourth Amendment question for the appellate courts.
• Retention Cut After the Lawsuit: what a city changed under litigation pressure, which is the practical measure of what these suits achieve.
• Plate Readers in Retail Parking Lots: the class action over private commercial deployments sharing data with federal agencies.
• Colorado's Warrant Requirement: a bill that would put older plate data behind a warrant.
• Kentucky's 90-Day Retention Cap: the deletion-deadline approach to the same problem.
• Washington's Driver Privacy Act: an attempt to bar immigration access outright and shorten retention sharply.
• A Bill That Died in a Tied Vote: the counterexample, and a reminder that most of these proposals fail.
• States Closing the Records Door: the countermove, exempting ALPR data from public-records law so audits become impossible.

7. The Sensors Alongside the Plate Cameras

Plate reading is one product line. The same vendor and the same contracts increasingly bring other sensors to the same poles.

• AI Audio Detection on City Streets: the expansion from gunshot detection into listening for human distress, and the Fourth Amendment questions it raises.
• Cameras and Microphones on a University Campus: what a single institutional deployment contains, installed with little notice.
• Doorbell Access Rebuilt Through Partners: the return of police access to consumer cameras by a different route after the direct deal collapsed.

Sources

1. OpenStreetMap taginfo, tag statistics for surveillance:type=ALPR. 138,351 objects worldwide, retrieved 9 August 2026.
2. Geofabrik regional taginfo, United States extract, tag statistics for surveillance:type=ALPR. 127,878 objects, retrieved 9 August 2026.
3. Electronic Frontier Foundation, What We Learned, review of ALPR data from 173 agencies covering roughly 2.5 billion scans in 2016 and 2017.
4. Wikidata, Flock Safety (Q108485435). American video surveillance company, headquarters Atlanta, inception February 2018. Retrieved 9 August 2026.

Related Guides

• Car Data Opt-Out Guide: the data your own vehicle reports, which is a separate stream from plate cameras.
• The Federal Car Surveillance Mandate: hardware requirements arriving in new vehicles.
• Know Your Rights in an ICE Encounter: the practical companion to section 3.
• Private Surveillance Networks and Police: the wider pattern of privately owned cameras feeding public agencies.
• The full guides hub.