New Hanover Redacted Who Searched Its Flock Cameras — and We Recovered the Searching Agency for About 96%

New Hanover County’s Sheriff’s Office runs a license-plate surveillance network, and when it finally produced its search audit, the file was enormous — roughly 2.98 million search records over about sixteen months. On every one of those lines, the county redacted the single field that answers the only question most residents would ask: who was doing the searching.

We rebuilt it. Using public records released by other states, we recovered the searching agency on the large majority of those searches. This is a companion to our report on one especially troubling use of this same system — searches the office labeled “Protest”. This piece is about the whole picture the redactions hid: who reaches into New Hanover’s cameras, how often, and from where.

One boundary first. This report is only about the Sheriff’s Office’s Flock license-plate cameras. It is not about the separate — and far larger — network of 2,147 private and business cameras the Sheriff has linked through its Axon Fusus real-time crime center, which we cover here. Every number below concerns Flock plate-reader searches alone.

The short version

• New Hanover’s cameras were searched by more than 4,600 distinct searching organizations in the recovered data.
• About 94% of that search activity came from outside North Carolina.
• The heaviest single searcher was Houston, Texas — more than 75,000 searches, from roughly 1,100 miles away. Dallas ran more than 55,000. Fairfax County (VA), Jacksonville (FL), Louisville (KY), and Cobb County (GA) each ran tens of thousands more.
• By state, Georgia, Texas, and Florida each out-searched North Carolina against New Hanover’s own cameras.

Data collected from New Hanover’s cameras was queried overwhelmingly by distant, out-of-state police agencies — and the county’s production concealed who they were. (These are searches that included New Hanover’s camera network, not necessarily searches aimed at a specific New Hanover resident.)

What surprised us most. Three things: nearly all of the search activity came from outside North Carolina; a single out-of-state department — Houston — accounted for more than 75,000 searches; and the searcher identities the county redacted could be reconstructed almost entirely from records other governments had already released.

What New Hanover produced, and what it hid

North Carolina law makes captured plate data confidential (N.C. Gen. Stat. § 20-183.32). But whether that confidentiality extends to every field in an administrative audit — including which government agency conducted a search — is a separate question, and other North Carolina agencies have answered it differently. Peer agencies in this state released the searching-agency field on the same platform, under the same law, within days of being asked. Whiteville produced roughly four months of it. New Hanover redacted the searching organization on all ~2.98 million lines, citing the confidentiality and criminal-investigation exemptions.

The result was a spreadsheet that documented an enormous volume of surveillance while revealing nothing about who was responsible for it. You could see that your county’s cameras were searched about three million times. You could not learn that Houston did roughly 75,000 of them.

How we rebuilt it

A Flock search does not stay inside one agency. It reaches across every camera network the searcher can access — often thousands of agencies at once — so the same search is logged in every jurisdiction whose cameras it touched. North Carolina redacts the searcher; many other states do not. So a search New Hanover hid from us often sits, intact, in the public records of an agency elsewhere that released everything.

We built a fingerprint for each New Hanover search from the fields the county left intact — the exact timestamp (to the second), the number of networks the search touched, and the search type — and matched those fingerprints against a large library of unredacted audits we obtained from public-records-friendly states. Where a fingerprint matched, the out-of-state copy supplied the searching agency the county had blacked out. The method is, in substance, the one USA TODAY used in its national Flock investigation — analyzing publicly released audit logs to see who searched whom — which we applied to the Cape Fear region. It is arithmetic on public records, not a hack or a leak.

An important caveat about the future: this kind of reconstruction will not be possible on searches going forward. Flock acknowledges that beginning in December 2025 it masked sensitive fields in its cross-agency (network) audits and introduced a standardized offense type; the company says the change protects officer safety and active investigations. The practical consequence is that future public records will carry less of what outsiders need to reconstruct searches like these. We were able to rebuild New Hanover’s redactions only because these searches predate that change and the out-of-state copies still carried the fields. That closing window is the subject of our report, the records that caught police abusing Flock are going dark.

We recovered a searching organization for the large majority of New Hanover’s audit — on the order of 96%. And we can quantify the method’s precision: of the roughly 2.82 million searches we recovered, 99.0% resolved to a single searching agency, and only 1.0% matched more than one agency on the fingerprint (those we treat as ambiguous). Two honest limits remain: a fingerprint match can occasionally be a coincidence — two agencies running the same search type at the same second would share a fingerprint — so the recovery figure is an estimate, not an exact count; and recovery depends on overlap, since a New Hanover search is recoverable only if it also touched a camera in a state that released its records. We treat the 1.0% ambiguous matches as unresolved rather than assigning them to any agency; they are too few to change the top-line findings. Any reporter or data editor can reproduce the check by matching New Hanover’s own fields against the out-of-state audits, which are public.

Who searched New Hanover’s cameras

The recovered searchers are overwhelmingly out of state. Ranked by number of searches against New Hanover’s cameras, the leaders are Houston (TX), Dallas (TX), Fairfax County (VA), Jacksonville (FL), Louisville (KY), and Cobb County (GA). By state, Georgia (about 12%), Texas (about 11%), and Florida (about 10%) each account for a larger share than North Carolina (about 6%).

One caveat matters for honesty: the “more than 4,600” figure counts distinct searching-organization names in the recovered data. We checked the obvious inflation risk — the same agency appearing as “Houston PD” and “Houston Police Dept” — and found effectively none, because Flock’s organization-name field is standardized; normalizing aliases did not meaningfully change the count. It may still include a small number of related sub-entities, so read it as a close estimate of national reach rather than an exact agency census. A ranked list of the top 100 searching agencies is available as a downloadable CSV, and the full tallies are available for verification.

Why redact it — and why the redaction fails

In practice, New Hanover’s redactions did not keep the searching agencies secret. Copies of the same searches sit in public records released by agencies in states whose disclosure laws differ from North Carolina’s — so the field the county blacked out here was, at the same moment, public in Texas and Georgia. Nothing in this report establishes that New Hanover’s redactions violated the Public Records Act. But they were not inevitable: other North Carolina agencies operating under the same ALPR statute released the searching-agency field. Whatever the rationale for the redactions, they did not actually keep the information secret. They kept it out of the records available to New Hanover residents.

New Hanover is also not unusual. When we ran the same reconstruction on the network audits other North Carolina agencies released, the national picture came back nearly identical — a heavily out-of-state searcher base led by the same distant departments, Houston chief among them. Our statewide finding, Who Is Searching North Carolina’s Flock Cameras, lays out that result across multiple agencies; New Hanover’s redacted audit, rebuilt, matches it closely. In other words, we did not need New Hanover’s cooperation to know roughly what its numbers would show — the rest of the state already showed us.

What Flock says an audit should show. Flock states that every query is logged and linked to the person who performed it, and it describes those audit records as the way to determine whether access was made by authorized people for legitimate purposes. Yet the records New Hanover released to the public name the operator on only a small fraction of searches. That does not necessarily mean Flock failed to record the identity — it most likely means New Hanover redacted it before producing the records.

New Hanover’s own searches — and its own redactions

Turn the lens inward, to searches New Hanover’s own personnel ran. Here the record is thinner, because the county redacted it again.

New Hanover’s outbound searches number about 76,654 between November 2024 and September 2026, across 85 operator identities. But the county redacted the operator’s name on roughly 79% of them — about 60,000 searches with no identifiable person attached. Of the fifth that show an operator, most appear only as an anonymous system ID; just 35 carry even a partial name. A case number appears on only about 6% of searches — and the county has said that is by design. Answering Commissioner Stephanie Walker, Sheriff McMahon wrote that “our policy does not currently require every search to be tied to a specific case number,” only a “purpose code.” The purpose code is the same self-entered field examined below. Separately, roughly 13% of operator accounts show round-the-clock search activity — the pattern you would get from either credential sharing or unassigned, role-based logins: a shared terminal, a dispatch or real-time-crime-center account, or 24/7 rotating shift work. We cannot tell which from the data, and we do not assert that credentials were shared. What every one of those explanations has in common is that it prevents a search from being tied to an accountable individual.

The reason field is no more reliable. Nearly a quarter of New Hanover’s own searches — about 23.6% — carry no reason at all. Most of the rest fall into a handful of broad, self-entered labels (“Drugs/Narcotics,” several spellings of “Investigation,” “Larceny/Theft,” “Other”), typed inconsistently. When we flagged that a number of searches were coded “traffic” — and North Carolina law expressly provides that plate-reader data “shall not be used for the enforcement of traffic violations” (N.C. Gen. Stat. § 20-183.31(b)) — the Sheriff’s Office said those codes were a miscoding. Either some searches were actually conducted for traffic enforcement, which state law prohibits, or the stated purpose codes were inaccurate, as the Sheriff’s Office says. Either possibility is an oversight problem: the audit field alone cannot reliably establish why a search occurred.

Add it up and the effect is this: whatever internal oversight the Sheriff’s Office conducts, the records it released to the public often do not let an outsider determine who ran a particular search, what case it concerned, or whether the stated purpose was accurate.

Why this matters

New Hanover County stood up a mass-surveillance network without a public vote, connected it to a nationwide system, and then produced records that hid who was using it — inbound and outbound alike. When journalists and residents across the country began using public audit logs to check for abuse, the response here was not more transparency; the county’s production showed less. The searchers we recovered, the county could simply have released — as its neighbors did.

Transparency is a choice, and other agencies have made it. After DeFlockILM reported that Appalachian State’s Flock cameras were feeding a nationwide network, the University switched off nationwide sharing; the Boone Police Department did the same after a Town Council resolution. New Hanover was asked to do likewise and declined.

None of this rests on trusting us. It rests on New Hanover’s own timestamps, joined to other governments’ lawful disclosures. The county released a system audit that kept the public from seeing who conducted millions of searches. We reconstructed much of what was redacted anyway.

Frequently asked questions

How many times were New Hanover County’s Flock cameras searched? The Sheriff’s Office produced roughly 2.98 million Flock network search records covering about sixteen months. A single Flock search can query cameras belonging to many agencies, so these records do not represent 2.98 million separate investigations.

How many searches per day did New Hanover’s cameras receive, and who ran them? Over the roughly 16-month window we reconstructed, New Hanover’s cameras were queried on the order of 5,700 times a day. Only about 13 of those per day were run by the New Hanover County Sheriff’s Office itself; the other ~5,650 a day — more than 99% — came from other agencies, about 94% of them outside North Carolina. (These count searches that included New Hanover’s camera network.)

Who searched New Hanover County’s Flock cameras? Our reconstruction recovered the searching organization for about 96% of the redacted audit. About 94% of recovered search activity came from outside North Carolina.

Which outside departments searched the most? In the recovered records, Houston, Texas accounted for more than 75,000 searches that included New Hanover’s camera network (not necessarily aimed at New Hanover residents), and Dallas more than 55,000. Other leaders included Fairfax County, Virginia; Jacksonville, Florida; Louisville, Kentucky; and Cobb County, Georgia.

Why could out-of-state police search New Hanover’s cameras? Flock lets agencies share access to license-plate-reader networks. One query can reach cameras operated by many participating agencies, depending on the sharing relationships in place.

Did New Hanover identify the agencies searching its cameras? No. In the records produced to us, the county redacted the searching-organization field. We reconstructed it by matching New Hanover’s remaining audit fields to unredacted public records released by agencies elsewhere.

Does a Flock “search” mean an officer searched a specific New Hanover resident? No. It means a query of the Flock system included New Hanover’s camera network. One query can search many participating networks at once, and these records document searches, not every camera detection.

How many Flock searches did New Hanover personnel themselves run? About 76,654 outbound searches between November 2024 and September 2026. The operator’s name was redacted on roughly 79%, a case number appeared on about 6%, and about 23.6% carried no reason in the field analyzed.

Can North Carolina police use plate readers to enforce traffic violations? No. State law provides that ALPR data “shall not be used for the enforcement of traffic violations” (N.C. Gen. Stat. § 20-183.31(b)). Searches labeled traffic-related therefore warrant explanation; the Sheriff’s Office has said such entries were a miscoding.

Is license-plate-reader data public in North Carolina? Captured plate data is confidential under N.C. Gen. Stat. § 20-183.32 and is not a public record. Whether every administrative field in an audit — including the identity of the government agency conducting a search — falls within that confidentiality is a separate question; other North Carolina agencies have released the searching-agency field.

Did New Hanover County taxpayers fund a surveillance system used mostly by other states? In effect, yes. New Hanover County pays for its Flock cameras — a roughly $219,000 Sheriff’s Office contract — yet about 94% of the recovered search activity against those county-funded cameras came from agencies outside North Carolina, led by Houston and Dallas, Texas. Local taxpayers built and maintain the network; out-of-state departments do most of the searching.

Does North Carolina law enforcement use Flock cameras and other ALPR to monitor protests? Yes, according to the Sheriff’s Office’s own records. Its Flock audit contains searches recorded under a case labeled “Protest,” with the reason “Intel,” on the morning of November 22, 2025 — hours before a protest at the Sheriff’s Office — and one of the searched plates belongs to an organizer of that protest. We document that in a companion report. Nationally, a leaked Flock training video and the Electronic Frontier Foundation’s review of public records show the same use by other agencies.

Can this analysis be independently verified? Yes. The reconstruction relies on timestamps and other fields in New Hanover’s own records matched against audit records released by other governments. Of the recovered searches, 99.0% resolved to a single searching agency. The methodology and limits are described below, and the underlying records and organization-level tallies can be made available.

Methodology, in full

For readers and reporters who want to check the work.

Source records. New Hanover County Sheriff’s Office Flock search/network audit, produced 2026 (~2.98M rows), searching organization, officer, plate, search ID, and case number redacted; timestamp, network count, time frame, search type, and offense/reason retained. Comparison library: unredacted Flock network and organization audits obtained via public-records requests from agencies in Washington, Georgia, California, and other states that release the searcher (and often the plate); plus HaveIBeenFlocked’s records, keyed to a consistent per-plate hash.

Fingerprint join. Each search was keyed on the exact search timestamp (UTC, to the second), the total networks searched, and the search type; New Hanover’s redacted rows were matched to the out-of-state library on that key, and matches supplied the redacted searcher. De-duplication used each search’s unique identifier where available and the fingerprint otherwise. Recovery: about 96% of the audit, of which 99.0% resolved to a single searching agency; a minority of fingerprint matches may be coincidental, so the figure is an estimate.

Counting organizations. The “more than 4,600 organizations” and “about 94% out of state” figures count distinct searching-organization names after removing exact duplicates; because a single agency can appear under several name variants, that count can overstate the number of truly distinct agencies, so it is best read as an estimate of broad national reach rather than a precise agency census.

Outbound figures. Operator, case-number, reason-code, and round-the-clock-activity statistics describe New Hanover’s own organization audit as produced. The round-the-clock-activity indicator counts accounts whose searches span most of the 24-hour clock; that pattern is consistent with credential sharing, with role-based or shared-terminal logins, or with 24/7 shift work, and we do not assert which — only that all of them defeat individual accountability. All figures describe the searches an operator ran, not every camera detection.

General information for public discussion, not legal advice.