One Search, a Thousand Networks

You share the number, and within hours someone pushes back.

When we reported that NC A&T’s sixteen campus cameras were searched 1,390,776 times in three months, the reply came fast: that figure is misleading. Those agencies in Texas and Florida weren’t hunting Greensboro drivers. They ran a search across the whole Flock network, and the system counted it against every camera group it touched.

Here’s the thing. They’re right. And once you understand why they’re right, that objection stops being a problem for you — because the honest explanation is a heavier charge than the headline ever was.

We read Flock’s own words — including the parts that argue against a scary number

We’re your neighbors in New Hanover County, and we’d rather get this exactly right than win a moment with a number we can’t defend. So we read Flock’s own documentation and the raw audit files line by line. Here is the precise version.

When a Flock search runs, Flock says the search — and the reason the officer typed — is written into the audit trail of every agency network whose cameras were included in that search’s scope. That is Flock’s own description of how the logs work.

So 1.4 million audit entries do not mean 1.4 million investigations aimed at that campus. They don’t mean 1.4 million officers chose NC A&T, or 1.4 million campus plates were pulled up, or 1.4 million matches were found. They mean roughly 1.4 million search events that included the university’s network within their scope. That is a different measurement — and we’ll say so plainly, because a guide you can’t trust is no use to you.

Why would a Houston detective’s search reach into Greensboro at all? Because of two contract settings: statewide lookup and national lookup. When they’re on — and reporting around the country shows they’re frequently on by default, sometimes without the buying city’s knowledge — the local cameras join a shared, searchable pool. In Mountain View, California, an audit found statewide lookup active on 29 of the city’s 30 cameras without the police department’s awareness; more than 250 outside agencies had run about 600,000 searches.

Say it precisely — “every camera in America” is the one line critics can knock down

It is tempting to shorten all of this to “one officer, one click, every camera in America.” Don’t. That’s the version a Flock spokesperson can rebut in a sentence.

Flock’s nationwide lookup applies only to agencies that opted in, is built around a complete, known plate, doesn’t necessarily include every Flock camera, and is switched off or restricted in some states. Partial-plate and vehicle-description searches — “gray pickup, roof rack” — generally run within the networks an agency is directly sharing with, not automatically across the whole national pool. The accurate, unshakable version is this:

One broad search can query thousands of participating agency networks across the country — and North Carolina’s were in the pool.

Still alarming. Just no longer rebuttable.

What an audit entry proves — and what it doesn’t

This is where precision becomes power. An audit entry proves one thing: that a network was included in a query. It does not prove that a local camera had ever seen the searched plate, that any local image was returned, that an officer opened a local result, or that the search broke any law.

There are really four different numbers hiding inside “1.4 million,” and they are not the same:

• Exposure — how often our network was included in a search.
• Result — how often our records actually matched.
• Access — how often someone viewed or exported our data.
• Misuse — how many searches broke a law or a policy.

The audit logs mostly show the first one. Anyone who tells you the number proves the other three is doing exactly what they accuse us of — reading more into the log than the log says.

So here is the version we will stand behind:

The totals are not millions of investigations targeting North Carolina. They are millions of outside search events in which North Carolina’s camera networks were included, because someone made those networks available through Flock’s sharing settings. The logs generally don’t tell us whether an NC record matched, was viewed, or was used — but they establish the scale of outside access those sharing decisions permitted.

That’s the whole case. It doesn’t need inflation.

Why the honest version still lands hard

Understanding the mechanism doesn’t soften the local findings. It sharpens them.

Our data is in the pool, and nobody here voted on that. Someone — the county, the university, or Flock’s default settings — opted North Carolina plate data into nationwide sharing. Under state law, disclosing captured plate data to another agency appears to require a written request (N.C. Gen. Stat. Ch. 20, Art. 3D, § 20-183.32; the request may be electronic). It is genuinely unclear whether Flock’s automated query workflow satisfies that — especially where no individualized request ever reaches the North Carolina custodian and no North Carolina record comes back. We asked New Hanover County for the paperwork authorizing this sharing; it has not been produced. (The statute’s separate requirement — a sworn, detailed statement to preserve specific data beyond the retention period — is a different mechanism; here we mean the everyday disclosure rule.)

The reasons still matter — carefully stated. At least 974 externally-initiated searches that included A&T’s network were labeled with immigration-related reasons. The logs alone don’t establish that any A&T record matched, or that anyone viewed one — but they do show the network was made available for a purpose the campus community never signed up for. Dayton, Ohio found more than 7,100 immigration-related searches of its cameras despite a city policy against exactly that. (More on federal and out-of-state access.)

The counting cuts both ways. If the objection is “most of those searches never returned a North Carolina image,” that concedes the system generates millions of audit entries that essentially no one reviews — which is precisely how officers in Georgia used these cameras to track ex-partners, and were caught only when somebody finally read the logs.

The Constitution is moving in one direction. In Carpenter (2018), the U.S. Supreme Court held that obtaining a week of a person’s historical cell-phone location data was a Fourth Amendment search. In Chatrie (June 29, 2026), it held that acquiring someone’s Google location history through a geofence request was a search too — and notably, in Chatrie the government actually had a warrant, and the Court still treated the sweep of innocent third parties’ location data as a serious constitutional intrusion. Neither case has yet held that a Flock query requires a warrant; applying them to ALPRs takes more legal work. But both strengthen the argument that government access to aggregated digital location histories can be a search — and a system whose default is a standing, nationwide search of everyone’s movements is exactly the kind of aggregation those cases worried about. What the logs do show is narrower and still striking: Flock’s workflow lets an officer run these searches without having to show that a judge signed off first.

Your plan is three steps

Start with the goal, because it’s bigger than a setting. The objection here isn’t only that our data is shared too far — it’s that a government has no business keeping a standing, warrantless record of where everyone drives in the first place. Mass surveillance of people suspected of nothing runs against everything the Fourth Amendment was built to protect; it is, plainly, un-American. Ending it is the destination. Turning off broad sharing is just the fastest first step on the way there. You don’t need to master the architecture to help — you need to do three things.

• See the mechanism — and describe it precisely. Not “every camera in America.” One broad search queries thousands of participating networks, and ours were in the pool. That sentence wins the argument because it can’t be knocked down.
• Start by turning off broad sharing — then keep going. Columbus, Ohio ordered its police to stop statewide sharing after an audit — a single administrative change that ends dragnet participation without removing a camera, and a fair first step any North Carolina agency can take this week. But shutting off the nationwide pool is harm reduction, not the finish line. The camera on the pole is still a warrantless government sensor logging every car that passes, and the audit logs, hot-list alerts, and one-to-one sharing deals still need daylight. The destination is ending the mass surveillance itself — and the county’s own contract lets it cancel at any time.
• Ask your county and your campus — out loud, on the record. Who authorized nationwide sharing of our plate data? What written authorization exists for it? And will you switch to vetted, one-to-one sharing? (Email your leaders and ask.)

What we lose if we stay quiet — and what we protect if we don’t

Stay quiet, and the default holds. Your movements sit in a pool that thousands of agencies can search, for reasons you will never see, with logs that no one in North Carolina is reading. The “they weren’t targeting you” defense becomes permanent — because they don’t target you. They search everyone, everywhere, every time, and you are simply always in range.

Speak up, and it ends differently. The first win is fast: the county flips the setting, publishes its sharing list, and pulls North Carolina data out of a nationwide pool — something an administrator can do this week, in public. But that’s the floor, not the goal. The goal is a county that doesn’t run a warrantless dragnet on its own residents at all — that decides, as a matter of principle, that free people moving through public space are not a database to be searched. We wrote the resolution so your board doesn’t have to start from scratch, and the contract can be canceled at any time.

August 17 — the date to put it on the record

The New Hanover County Board of Commissioners meets August 17, 2026. Public comment is three minutes — enough to put it on the permanent record: pull us out of the nationwide pool now, and set a course to end the warrantless surveillance for good. (How to speak — and what to say.)

Sources

• Flock Safety — statement on network sharing, use cases, and federal cooperation (how searches and audit entries work).
• DeFlockILM — the NC A&T Flock audit (1.4 million searches), with the full dataset in our records repository.
• WOSU Public Media — Columbus orders police to stop statewide Flock sharing after an audit (July 11, 2026).
• N.C. General Assembly — Chapter 20, Article 3D (ALPR statute).
• U.S. Supreme Court — Chatrie v. United States (June 29, 2026).

DeFlockILM is a citizen effort on ALPR and privacy issues in the Wilmington area. The audit files cited here are unaltered public records, available in full in our NC Records Repository. Our description of Flock’s architecture is drawn from Flock’s own published statements and the audit logs; if we have any mechanism wrong, tell us — we correct our work in public. General information for public discussion, not legal advice. Last updated July 13, 2026.