No Mass Surveillance of Ordinary People

What ALPRs Capture

Every Car. All Day. Suspected of Nothing.

An automated license plate reader (ALPR) is a camera paired with software that photographs every vehicle passing its field of view and extracts a record from each photo. A typical record includes:

• the license plate number;
• the camera's location;
• the date and time of the scan;
• a photograph of the vehicle; and
• often a "vehicle fingerprint": make, model, color, and identifying details such as roof racks, bumper stickers, or damage.

The scan happens for every car, all day, whether or not anyone is suspected of anything. The overwhelming majority of records describe vehicles that match no watch list at all - ordinary people, never accused of any crime, going about their lives.

Networks & Sharing

The Camera Is Local. The Database Is Not.

Modern ALPR products, including Flock Safety's, are sold as networks. Each scan is uploaded to a database hosted by the vendor, not kept in a filing cabinet at the local police department. Depending on contract settings, that database can be searched by:

• the agency that installed the camera;
• other agencies in the same network, sometimes across the country; and
• in some configurations, federal agencies or task forces.

This is the part most residents never hear at the council meeting: a camera approved to find stolen cars in one town can feed a nationwide, vendor-hosted search index. Who can query it, for what purpose, and with what audit trail is governed by contract language and policy - documents that are often not published until residents demand them.

Sharing settings, retention periods, and audit logging vary by jurisdiction and can change after the vote that approved the cameras. That is why our interim safeguards demand published contracts, policies, and audit logs.

The Aggregation Problem

Why the Sum Is the Danger

One scan is a moment. A network is a dossier.

A single scan is a small fact: one vehicle, at one place, at one moment. Standing alone it reveals almost nothing. The problem is what happens when thousands of scans from a networked system are combined - aggregated - and made searchable by plate number.

Query a plate across a network's history and the scans line up into a timeline. The timeline reveals patterns, and the patterns reveal a life:

• where a person works (the same block, every weekday morning);
• where they worship (the same address, every Sunday or Friday or Saturday);
• where they seek medical or mental health care (a clinic parking lot, repeatedly);
• which political events they attend, and where they assemble or protest;
• whose homes they visit, and when; and
• when and where they travel, and when their car is not at home.

Nobody decided to tell the government any of that. It falls out of the mathematics of collection. This is the same aggregation logic the Supreme Court confronted with cell-site records in Carpenter and location history in Chatrie - see Records: Landmark Records & Cases.

The Money

More Officers. Fewer Cameras.

These systems are sold as subscriptions, and the meter never stops. Industry pricing guides put Flock's cameras at roughly $2,500-$3,000 per camera per year, plus one-time install fees - and a 2024 price increase added about $500 per camera per year on new contracts, with renewal rates not guaranteed (pricing guide; ballpark figures from industry sources, so prefer the contract numbers below).

The actual contracts say it plainly. Denver's rejected renewal was $666,000 for two years; the mayor then signed a $498,500 extension anyway (Denverite). Sedona's canceled contract carried a first-year invoice of about $51,146 covering 12 cameras - 11 cameras were installed and removed (KNAU). Oklahoma City pays $270,000 per year for 90 cameras (KFOR).

And the "free pilot" is a hook, not a gift. Gig Harbor, Washington was offered cameras with a state grant covering installation and the first year - and the council still said no, because the city would be on the hook every year after (Gig Harbor Now).

Every one of those dollars is a public dollar, spent year after year on a subscription. We would rather see public money put officers on the street than build camera networks that track everybody. Recurring six-figure subscriptions, unilateral price hikes, and a database of everyone's movements: that is not free. It is a permanent line item purchased with both dollars and liberty.

Glossary

Terms, Plainly Defined

ALPR (automated license plate reader)

A camera-plus-software system that photographs passing vehicles and extracts plate numbers and vehicle details into a searchable database record.

Hit

A scan that matches a watch list, such as a stolen-vehicle or wanted-person list. Hits are the tiny fraction of scans the systems are marketed on.

Non-hit data

Every other scan: records of vehicles matching no list at all - ordinary people, going about their lives. This is the bulk of every ALPR database, and its retention is the heart of the controversy.

Retention

How long scan data is stored before deletion. Every additional day of retention deepens the travel history that can be reconstructed. We ask that non-hit data be deleted within 48 hours or less.

Geofence warrant

A court order compelling a company to identify devices or accounts that were within a drawn area during a time window - searching outward from a place, rather than from a suspect. The Supreme Court's Chatrie decision (2026) held that acquiring Google Location History this way is a Fourth Amendment search.

Aggregate

To collect many small pieces of information into one larger whole. Each scan is minor on its own; aggregated, scans become a searchable record of a person's movements over weeks, months, or years.

Saying It Accurately

What the Law Does and Does Not Say

We hold ourselves to a legal accuracy rule: we never claim that "every ALPR is already unconstitutional." No court has held that. What the Supreme Court's aggregation cases establish is that assembling detailed location histories raises grave Fourth Amendment concerns and demands strict constitutional scrutiny. That is the accurate claim, it is fully sourced, and it is more than enough reason for a community to say no.

Case-by-case detail, with citations and official links: Records: Landmark Records & Cases.

Heard at Every Council Meeting

Answering the Objections

The same objections come up every time a community debates these cameras. Here is each one, answered calmly, with only verified and sourced evidence.

Objection One: "But It Catches Criminals"

Sometimes it does. The honest question is how often, at what error rate, and at what cost to everyone who is not a criminal.

• Error rates are real. The Los Angeles Police Department's Inspector General audited the department's ALPR program - multiple vendors, not Flock alone - and its counts (161 false alerts of 498 reviewed) work out to a 32.3 percent false-positive rate on stolen-vehicle alerts - roughly one bad alert in three (ABC7). LAPD did not renew, citing civil-liberties concerns and unresolved data-ownership terms.
• Oversight has been absent at scale. The California State Auditor's Report 2019-118 (Feb 2020) found LAPD holding roughly 320 million plate images with no ALPR-specific policy, audited agencies lacking access and monitoring rules, and data shared with entities in 44 of 50 states (official report, PDF; EFF summary).
• Independent efficacy research is thin and mixed. A study of Atlantic City's ALPR expansion found no reduction in violent crime, and the associations it did find (shootings, vehicle theft) were hard to separate from nationwide crime declines (Shjarback & Sarkos, Justice Evaluation Journal (2025); CNN overview). We do not claim ALPRs never help solve crimes - that would be its own overclaim. The accurate statement is that independent evidence of crime reduction is thin and mixed.
• The impressive numbers are vendor-funded. The widely repeated claims that Flock "helps solve 10 percent of reported crime" and raises clearance rates come from research Flock itself commissioned, surveying its own customers, without selection-bias controls. They are vendor-funded claims, not independent evidence (Police1 coverage of the vendor study).

Objection Two: "The Data Is Only Used Properly"

The record says otherwise, and the pattern is consistent: the most common documented abuse is officers tracking women they know.

• Kechi, Kansas. A police lieutenant used Flock and Wichita police cameras to track his estranged wife - finding locations she had deliberately hidden from him, including her children's day care. He was arrested in October 2022, pleaded guilty to computer crime and stalking, received 18 months of probation, and had his Kansas police certification revoked (reporting; follow-up).
• Riverside County, California. A former sheriff's deputy was convicted on 10 charges including stalking and unauthorized data access against his former fiancee, and sentenced in February 2026 to six years in prison (Marshall Project roundup).
• The scale. CNN (July 2026) reported at least two dozen cases of officers misusing Flock to track people, most often romantic partners, citing the Institute for Justice - whose tally has since grown to dozens (82 incidents). The Washington Post (August 2026) counted at least 50 officers accused or charged with misusing Flock or similar ALPR systems (CNN; Institute for Justice).
• Here at home. An Oklahoma City police sergeant, Sgt. Keenan Heskett, was reprimanded, given a three-day unpaid suspension, and placed on a year of disciplinary probation for using an official police database for personal use - a sanction disclosed at the August 18, 2026 contract-renewal meeting (KGOU).

Every one of these searches was made by a credentialed insider. No hack was required - only a login and a bad motive.

Objection Three: "It Stays Local"

The network is the product. "Local" cameras feed a nationally searchable database, and the documented record shows the data traveling far beyond the town that approved it.

• Federal side-door access, at scale. 404 Media's analysis of Flock audit logs found more than 4,000 nationwide or statewide lookups performed by local and state police for the federal government, as "informal favors," or with a possible immigration focus - warrantless access for ICE, which has no Flock contract of its own (404 Media). Congress opened an investigation in response (404 Media).
• Illinois: illegal access, officially found. The Illinois Secretary of State found Flock had illegally allowed U.S. Customs and Border Protection to access Illinois cameras - the finding that triggered the Oak Park and Evanston cancellations (Wednesday Journal).
• Dayton, Ohio. The city suspended its program and physically covered all 72 of its cameras after roughly 7,100 immigration-related searches violated city policy (Fortune).
• Washington State. After the UW Center for Human Rights documented Border Patrol access to Washington Flock data, Auburn's police department turned off federal-agency access in its system (UWCHR).
• The Texas search. On May 9, 2025, a Johnson County, Texas sheriff's deputy queried Flock's national network - about 83,000 cameras across roughly 6,800 networks, including in states where abortion is legal - with the logged search reason "had an abortion, search for female." The sheriff and Flock later characterized it as a welfare check on a woman whose family feared for her safety, not an attempted prosecution, and no charge against the woman is documented. The verified facts are stark enough: the search happened, it was warrantless, it was nationwide, and that is what the log said (404 Media; EFF; Snopes).

Oklahoma City's August 2026 renewal disabled Flock's automatic national sharing and gave the City case-by-case authority over federal sharing - a veto power won by public pressure, and proof the "it stays local" default was never the default. Details: The Honor Roll.

Objection Four: "The Courts Have Settled This - Nobody's Suing"

People are suing, and the courts have not settled it. We state this one with particular care, because our own side sometimes overclaims it - and we do not.

• Schmidt v. City of Norfolk (federal). Residents challenged Norfolk, Virginia's network of roughly 172-176 Flock cameras; the city's own data showed the plaintiffs' cars logged 526 and 849 times in about four months. On January 27, 2026, the district court granted summary judgment for the city, holding the network does not violate the Fourth Amendment "yet." The plaintiffs have appealed to the Fourth Circuit (No. 26-1227), pending as of August 2026 (WHRO; Courthouse News). Anyone who tells you a court ruled Flock unconstitutional in Norfolk has it backwards - so far.
• Virginia's state courts split, then reversed. A Norfolk Circuit Court judge suppressed warrantless Flock evidence, likening the citywide network to GPS tracking - but the Virginia Court of Appeals reversed in an unpublished opinion, holding plates in public view carry no reasonable expectation of privacy (GovTech; 13NewsNow). We cite both rulings because honesty about losses is the price of being believed about wins.
• The other direction. The Supreme Court's aggregation principle - Carpenter (2018) and Chatrie (2026) - holds that assembling a detailed record of a person's movements implicates the Fourth Amendment. How that principle applies to plate reader networks is exactly what is being litigated now. See Records: Landmark Records & Cases.

"The courts are still working through it" is not a reason to wait. It is the reason communities decide for themselves at the council table - which is where this campaign works.

The Bottom Line

Each objection above fails on the verified record. The remedy is not a better dashboard or a longer policy document. It is removal: take networked mass surveillance out of public spaces, and require constitutional process before government tracks anyone. Read the full ask.

Every claim in this section was verified against the linked sources on October 16, 2026. Our corrections policy - and how to tell us we got something wrong - is on Sources & Corrections.

Nationwide Tracker

Communities That Ended, Paused, or Refused Flock

How this tracker works

Action categories are labeled separately, never lumped together:

Canceled

An existing contract was terminated before its end date.

Non-renewed

A contract was allowed to expire; the jurisdiction declined to renew it.

Deactivated

Cameras were switched off or physically covered, possibly while a contract remains in force.

Suspended

Use was paused pending review; the program may resume.

Rejected

A proposed program was voted down or withdrawn before installation.

Restricted

The program continues, but with new binding limits on use, sharing, or retention.

Reversed

A program that had been ended or switched off was later turned back on. The row stays in the table so the record shows the reversal, not just the ending.

We never state a national total without a date, a link to the underlying list, and a definition of what is being counted. Every entry carries a last-verified date. Sources rank in this order: official records (ordinances, contracts, meeting minutes and votes); signed documents, audits, and meeting video; reliable local reporting; advocacy trackers, treated as leads only and verified independently before we publish. The recognition side of this record - the officials and departments who stood up - is on The Honor Roll.

The DeFlock Movement

See the Cameras. Map the Cameras.

Surveillance thrives on invisibility: most people have no idea a camera on a utility pole is logging their plate. The DeFlock movement answers that with transparency. DeFlock.org is a free, open-source, volunteer-built map of ALPR camera locations across the country, layered on OpenStreetMap data and grown by residents documenting cameras in their own towns. Locally, DeFlockOKC organizes the same work across the Oklahoma City metro.

You cannot debate a camera network you cannot see.

Use the map below: look up your own commute. Seeing how many times your plate is read on an ordinary Tuesday is the fastest way to understand the aggregation problem - and mapping is itself lawful, peaceful civic documentation: the cameras sit in plain view in public space, photographing you.

Map data: the volunteer-built DeFlock map (opens full-screen), built on OpenStreetMap. An embedded third-party map loads from deflock.org when this section is viewed; our own site still sets no cookies and runs no trackers.

Limiting your exposure, lawfully

Driving is public. You have no obligation to route your errands past surveillance cameras, and checking the map to choose streets without them is entirely lawful - like choosing not to walk through a metal detector you are free to walk around.

• Know your area. Check the map for camera placements on your regular routes; many cluster on main arterials and intersection chokepoints, and residential or secondary streets often have none.
• Add what you find. If you spot an unmapped camera, document it and submit it - the map only works because residents keep it current.
• Do not obscure your plate. Plate covers, sprays, flippers, and obstructions are illegal in Oklahoma and most states. This project does not endorse them, ever. The problem is not your plate; it is the network.

And the honest limit: individual avoidance is a stopgap, not a solution. You should not have to study a map to run errands unrecorded in a free country. The fix is not clever routing - it is removal. That is why this campaign exists.