Wylie Said No to Flock

organizations have been granted access to Wylie's Flock data (Flock's own portal, as of August 26, 2026). The list includes agencies in Oklahoma, Kansas, and Ohio, the Texas Department of Public Safety, the Texas Department of Criminal Justice, and even Union Pacific Railroad police. Wylie's cameras also pull from 829 more networks, 522 of them out of state, reaching into more than a dozen states, federal agencies (including the U.S. Postal Inspection Service and Wright-Patterson Air Force Base), private railroads, a national retailer, and even a homeowners association.

"It only reads license plates."

You will hear this from officials, and Flock's own paperwork contradicts it. The sales letter Flock sent Wylie, signed by its CEO, advertises machine vision that identifies vehicles by "roof rack, bumper stickers, etc." and captures cars "with the absence of a license plate." Flock even trademarked the profiling: it's called Vehicle Fingerprint®.

Wylie PD said the same thing to council on February 10: searches can run on a full plate, a partial plate, "or it could be put in for some form of descriptors." A system you can search without a plate is not a system that only reads plates. This is AI image analysis, in Flock's own words: the sales letter calls it "proprietary machine vision," the marketing calls the cameras "AI-powered," and Flock's LPR Policy admits it keeps residents' images "for the purpose of improving Flock Services through machine learning": your car's photos train their AI. Every passing car is analyzed and profiled: color, make, model, damage, stickers, racks. The plate is just one field in the profile, and the newest tool (OS Investigate, below) doesn't even need the car's description, just its driving pattern.

Sources: Flock Safety Sole Source Letter provided to the City of Wylie, 2024 (signed by CEO Garrett Langley); Wylie City Council work session transcript, February 10, 2026; details in Wylie's Documents below.

September 2026: WIRED Rebuilt Flock's AI Tool. It Searches for People, Not Just Plates.

For years, Flock told the public its cameras "cannot recognize, identify, or track individuals." On September 3, 2026, WIRED published the tool that does exactly that, reconstructed from Flock's own code sent to police officers' browsers. It is not a leak or a rumor. It is Flock's software.

What the tool can do, per WIRED:

• Search for people by written description. A feature Flock calls "FreeForm" lets an officer type a plain-language description of a person, clothing, colors, objects, even tattoos, and search video footage for anyone who matches. Flock says it has sold this for years.
• Run a continuous AI watch list. An officer can draw a shape on a city map and make every camera inside it run a nonstop, automated search for anyone fitting a description, alerting the officer when the AI thinks it has found a match.
• Infer who you associate with. Enter one license plate, and the tool hands police the other vehicles "seen beside it," building a picture of your associations from driving patterns.

The guardrails record, but do not stop. WIRED found the safeguards Flock announced may log and discourage abuse but will not prevent it. Some searches can currently run with no reason and no case number, and the automated audit tool is off by default. And the AI content filter blocks searches by race or religion, but treats "political, social, and cultural expression" as only a warning an officer can click through, the exact category the First Amendment protects most. Experts note one use of such a search is to see who attended a protest.

This is not hypothetical abuse. Per WIRED and the Washington Post, at least 50 U.S. officers have recently been charged with or accused of misusing plate readers, 46 of those cases involving Flock, and in more than half the target was a wife, girlfriend, ex-partner, or a woman the officer wanted to meet. When one South Carolina sheriff's office finally turned on the audit tool, an internal-affairs officer flagged more than 2,700 allegedly unauthorized searches the next day. Two members of Congress wrote to Flock's CEO after a Texas deputy searched more than 83,000 cameras for a woman who had an abortion.

Source: WIRED, "This Is Flock's AI Search Tool for Cops," September 3, 2026; officer-misuse figures via The Washington Post.

July 2026: Wylie's Own Congressman Says Flock Needs a Warrant Requirement

U.S. Rep. Keith Self (R-McKinney), who represents Wylie in Texas' 3rd Congressional District, introduced the PRIVACY Act in July 2026. His bill would require federal agencies to get a warrant before accessing Flock or any ALPR data, cap data retention at 30 days, create a federal registry of surveillance technology, and bar federal funds from paying for these systems.

Self's original cosponsors are Rep. Eli Crane (R-AZ) and Rep. Andrew Clyde (R-GA), who called Flock cameras part of "the federal government's dangerous web of warrantless surveillance." In Self's words: "Americans' Fourth Amendment right to be secure in their privacy does not disappear just because of new technology or artificial intelligence. This bill simply restores that protection."

It's worth being straight about where law enforcement stands: groups like the Texas Municipal Police Association support ALPR technology. That's exactly why a warrant requirement is the right ask rather than an argument about whether cameras are ever useful. A warrant doesn't take the tool away from police. It puts a judge between a database of residents' movements and anyone who wants to search it. When Wylie's own congressman says this technology needs constitutional protections it doesn't currently have, the question for City Council is simple: why is Wylie feeding a system its own representative considers a threat to Fourth Amendment rights?

And Self isn't alone. Days later, Rep. Tim Burchett (R-TN) introduced H.R. 9800, the Protection Against Mass Surveillance Act, which goes further: it would bar federal agencies from purchasing, operating, or accessing ALPR and biometric surveillance systems outright, bar federal funds from paying for them anywhere, and make improperly obtained data inadmissible in court. Both bills come from conservative Republicans, one of them Wylie's own congressman. The guardrails argument is no longer a left-versus-right fight.

And here's the part that made this website matter: neither bill touches a city-funded system like Wylie's. Congress can close the federal side-door, but only Wylie's council could close the front door, and on August 25, 2026, it did, with a unanimous 7-0 vote to end the contract.

Rep. Self took notice. After the vote, he shared the news of Wylie's decision publicly, writing that "local elected officials are listening to their constituents." The congressman who wants a warrant requirement at the federal level pointed to Wylie's council as an example of exactly that principle working at home.

Sources: The Texan, July 2026 (PRIVACY Act); H.R. 9800, 119th Congress (bill text); Police1 on H.R. 9800

August 2026: A U.S. Senator Opens a Congressional Investigation into Flock

On August 26, 2026, Sen. Josh Hawley (R-Mo.), who chairs the Senate Judiciary Subcommittee on Crime and Counterterrorism, opened a formal investigation into Flock Safety's collection, retention, and dissemination of the data its cameras generate. He sent a letter to Flock CEO Garrett Langley demanding answers about the company's data safeguards, with a response due by September 8, 2026.

In his own words: "Congress never authorized the network your industry has built." And: "Americans do not surrender their privacy rights when they drive to work, drop their kids off at school, or go to church. The overwhelming majority of the Americans captured in those records did nothing wrong."

This is the same concern Wylie residents raised, now at the highest level of federal oversight. Taken together, the pressure on Flock is genuinely bipartisan: House bills from Reps. Self and Burchett, warnings from Sen. Rand Paul and Sen. Bernie Sanders, a Texas order from Gov. Abbott to cut off state funding, and now a Senate investigation led by a Republican committee chairman.

Sources: Office of Sen. Josh Hawley, August 26, 2026; The Hill; Axios

August 2026: Texans Funded This Through a $1 "Catalytic Converter" Fee They Were Never Told Paid for Surveillance

In 2023, the Texas Legislature unanimously added $1 to every auto insurance policy in the state, sold to the public as a way to fight catalytic converter theft. That fee has since raised about $81 million.

A state board whose members are largely appointed by the governor, the Motor Vehicle Crime Prevention Authority, has quietly steered at least $30 million of it into more than 3,200 Flock cameras across Texas, plus another $15.9 million for a Department of Public Safety contract covering roughly 1,200 more. Money collected to fight car theft was routed into building a statewide surveillance network, without most Texans ever knowing.

Even a lawmaker who voted for the fee was stunned. Rep. Brian Harrison: "In a million years, I never could have even contemplated that this would be used to fund what is effectively warrantless surveillance."

This is the money Gov. Abbott has now ordered the state to stop spending on Flock. And it is exactly why Wylie declining to renew matters: it keeps our city out of a statewide expansion that residents were funding without their knowledge or consent.

Source: The Texas Tribune, via ABC13 Houston, August 30, 2026

June 2026: Flock Leaked Police Searches onto Google-Style Search Engines

In June 2026, 404 Media confirmed that Flock's own system exposed law enforcement license plate search queries to the public internet, indexed by DuckDuckGo and Bing. Anyone could find real police searches containing license plate numbers, vehicle make/model/color, identifying details like "window stickers," case numbers, search reasons, and date ranges. Reporters reproduced the searches themselves.

Flock acknowledged roughly 70 indexed results from 2024–2025 and said it was working to remove cached content. This is the company Wylie trusts with a record of everywhere you drive.

Source: 404 Media, June 11, 2026

License plates exposed in January 2026, when police departments nationwide released completely unredacted Flock audit logs through routine public records requests, revealing millions of surveillance targets, including people in active investigations. The data covered tens of millions of searches by thousands of departments.

Flock's response was not to fix the flaw. Instead, Flock severely restricted what its audit logs show, gutting the very transparency tool cities like Wylie point to as their accountability safeguard, and sent takedown demands the Electronic Frontier Foundation called "bogus."

Source: 404 Media, January 13, 2026

It Already Happened in Texas: 83,000 Cameras Searched Over an Abortion

In 2025, the Johnson County Sheriff (about an hour from Wylie) used Flock to run two nationwide searches for a woman who had self-managed an abortion: first across 17,684 cameras, then across 83,345 cameras in 6,809 networks. The search log entry read: "had an abortion, search for female."

When the story broke, the sheriff and Flock publicly claimed it was a "missing person" welfare check. Documents released to the Electronic Frontier Foundation, including a detective's sworn affidavit, showed it was formally a death investigation, that investigators consulted prosecutors about charging her, and that they reviewed her private text messages.

The DA confirmed no crime had been committed. She was tracked across the country anyway, through cameras in cities that never agreed to be part of a Texas abortion investigation. Cameras like Wylie's.

Flock's response when this broke was denial, not concern. The company's own blog post about the incident, signed by CEO Garrett Langley, titled a section "Correcting Misinformation" and called the reporting "purposefully misleading." In a Forbes interview the same year, Langley went further, defending the search itself on camera: "This is everything's working as it should be... due to the political climate, this was really good clickbait." The sheriff who ran the search has since been indicted on a felony perjury charge (the case is still pending, with no conviction as of August 2026), a fact that surfaced only because journalists, not Flock, kept digging.

This wasn't isolated. Illinois's Secretary of State separately found 46 out-of-state agencies had searched Illinois Flock data without the legally required attestation not to use it for abortion or immigration enforcement. One town's system alone was queried 262 times for immigration reasons in under four months. Reps. Raja Krishnamoorthi (D-IL) and Robert Garcia (D-CA) opened a formal congressional investigation in August 2025, citing a parallel case of an ATF analyst improperly querying Richmond, Virginia's network for immigration purposes despite the city's explicit non-cooperation policy. The same pattern (out-of-state agencies reaching into a city's data without its knowledge or consent) turned up in California too, where the state's Attorney General sued the city of El Cajon in October 2025 over plate data shared with more than a hundred out-of-state agencies. Flock has since had to pull California, Illinois, and Virginia from its national search tool specifically because of repeated violations in each. Every fix followed a journalist or a state audit catching the company first, never the other way around.

Sources: Electronic Frontier Foundation, October 2025; 404 Media (original report); Flock Safety blog, "Setting the Record Straight" (June 19, 2025); Forbes, "AI Startup Flock Thinks It Can Eliminate All Crime In America" (Sept 3, 2025: "terroristic organization" quote, "everything's working as it should be"); Forbes, "Flock CEO Apologizes For Calling Activists 'Terrorists'" (July 17, 2026); Capitol News Illinois (46 out-of-state agencies); Congressional investigation, Aug 6, 2025

Flock hot-list alerts that were FALSE, per the Los Angeles Police Department's own Inspector General audit (July 2026). Of 498 alerts reviewed, 161 were wrong: stale or incorrect database entries that the cameras faithfully matched and pushed to officers as real-time hits on supposedly wanted vehicles.

The day after the audit published, LAPD, the second-largest police department in America, let its Flock contract expire, citing "serious concerns around civil liberties... particularly around privacy and the data being collected."

The stakes of a false alert are not abstract. That same summer, a Wisconsin woman was pulled out of her car at gunpoint twice in one week, by two different departments, on the same erroneous Flock hit that police failed to clear even after the first stop was confirmed to be a mistake (FOX6 Milwaukee; full story in Real Victims below).

Sources: TechCrunch, July 13, 2026; Biometric Update (LAPD OIG audit)

Texas School Cameras Became an ICE Back Door

Audit logs from six Texas school districts show their Flock cameras are searchable through the national network by thousands of outside agencies. One Houston-area district's 8 school cameras were searched more than 733,000 times in a single month by over 3,100 agencies, with immigration-related queries outnumbering criminal ones two to one. The districts themselves never ran an immigration search; the network did it for them.

Meanwhile, the Texas Tribune reports TxDPS Flock partnerships with local agencies jumped 69% in six months (122 agencies in February 2026 to 206 by August), and confirms Texas agencies have run Flock searches on behalf of federal immigration agents. Recall Wylie PD Policy 906 mandates sharing Wylie's data with TxDPS.

Sources: The 74 investigation; Texas Tribune, August 3, 2026

From Bernie Sanders to Rand Paul: the Entire Spectrum, One Message

Sen. Bernie Sanders (I-VT), August 18, 2026, about Flock by name: "Going to the doctor? Flock knows. Dropping your kid off at school? Flock is tracking you." He called Flock's network mass surveillance and demanded it stop.

Sen. Rand Paul (R-KY), the same month, about the same company: "This isn't crime fighting. This is mass surveillance of every American, and a violation of our 4th Amendment rights."

Think about the last time Bernie Sanders and Rand Paul agreed on anything. When the furthest-left and furthest-right senators in the country describe one product in the same words, that isn't politics. That's the Constitution.

Sources: Sanders statement, August 18, 2026; Daily Caller, August 19, 2026 (Paul statement)

Knox County, Tennessee: Mayor Glenn Jacobs (Republican). In late July 2026, Jacobs called for a pause on his own county sheriff's plan to add roughly 270 new Flock cameras; installation was halted. On August 9 he went on national television to call for a nationwide pause on Flock deployment until there are constitutional guardrails. His framing could be Wylie's own:

"This is not about depriving [law enforcement] of the tools. It's about figuring out how these tools can be used in an appropriate manner that is consistent with the Constitution." — Mayor Glenn Jacobs

Connecticut: Governor Ned Lamont (Democrat). On August 7, 2026, Lamont asked Connecticut towns to immediately pause installing new cameras pending a statewide policy review, giving the state's police standards council 30 days to produce recommendations. A Democratic governor and a Republican mayor, same week, same ask: pause, then add guardrails. And on August 18, Rhode Island's Governor Dan McKee became the second governor to act, ordering a 60-day statewide review of license plate readers, pausing new state installations, and asking every municipality to freeze new cameras until uniform standards exist. (WPRI, August 18, 2026)

Sources: WVLT (Jacobs); Fox News (Jacobs, nationwide pause); CT.gov (Lamont)

Flock Has Never Been Proactive. Not Once.

In August 2026, under mounting pressure, Flock announced new "safeguards": cutting data retention from 30 days to 7, requiring case-number codes for searches, and mandating misuse audits. It's worth asking the obvious question: why weren't these there from the beginning? The answer is Flock's entire track record. Every privacy protection the company has ever added came after an abuse it was built to quiet, never before it.

• March 2025: Shut off national data-sharing for California, after audits caught out-of-state and federal agencies illegally querying California data.
• August 2025: Paused its federal CBP/ICE pilot programs, after reporting revealed Customs and Border Protection had backdoor access to 80,000+ cameras, and an Illinois state audit found the access violated state law.
• January 2026: Added a one-click "turn off federal sharing" toggle, after the immigration-access stories piled up.
• February 2026: Added warning disclaimers about abortion and immigration law, after a Texas sheriff used Flock to search 83,000 cameras for a woman who had had an abortion.
• Spring 2026: Restricted employee sales demos to "public areas" only, after Dunwoody, Georgia's own records showed Flock employees had viewed that city's camera feeds more than 1,000 times since 2023, including a children's gymnastics room camera used as a sales demo. Flock acknowledged the access while disputing that it was improper.
• April 2026: Made its misuse-detection feature available, after officers nationwide were caught using Flock to stalk women.
• August 2026: The retention cut and mandatory audits above, after a summer of officer-stalking resignations and a wave of cities canceling contracts, announced days before a national week of action.

None of it existed at the start. If a warrant requirement, shorter retention, and real audits are good ideas now, they were good ideas in 2021 when Flock began. For years, privacy advocates said exactly that and were waved off. None of this abuse was hard to predict. People warned about all of it. Flock built the surveillance network first and bolted on protection only when the outcry grew loud enough to threaten the business.

Why this mattered for the renewal decision: Flock's August "changes" don't take effect until January 1, 2027, after Wylie's renewal deadline, and after the contract would have locked the city in for two more years. Renewing wouldn't have bought the "new" Flock; it would have renewed the old one, betting that a company whose entire history is reactive would suddenly protect residents proactively. It never has yet. And even the revamped version still doesn't require a warrant, the one protection that actually matters. The Institute for Justice called the changes "window dressing" for exactly that reason. On August 25, the council agreed, and voted to walk away.

Sources: Associated Press (Flock platform changes, Aug 2026); Illinois Secretary of State audit; 404 Media (CBP pilot); Bexar County Flock transparency page (30-day retention); KERA News (Denton, 30-day retention); 404 Media (Dunwoody employee access); Wylie City Council work session transcript, February 10, 2026.

Governors

• Gov. Greg Abbott (R-TX) ordered all Texas state agencies to stop funding Flock cameras (August 28, 2026). Dallas is already deactivating 321 of its cameras as a direct result.
• Gov. Ron DeSantis (R-FL) ordered Flock cameras removed from Florida's state highways within 30 days, and the state revoked every license-plate-reader permit, with the governor warning against a "digital AI surveillance state." Florida counties began pulling their cameras within days.

United States Senators

• Sen. Josh Hawley (R-MO), a Judiciary subcommittee chair, opened a formal Senate investigation into Flock: "Congress never authorized the network your industry has built."
• Sen. Rand Paul (R-KY): "This isn't crime fighting. This is mass surveillance of every American, and a violation of our 4th Amendment rights."
• Sen. Bernie Sanders (I-VT), about Flock by name: "Going to the doctor? Flock knows. Dropping your kid off at school? Flock is tracking you."
• Sen. Ron Wyden (D-OR) pressed Flock into new privacy protections and is pushing federal investigations of surveillance-data sharing: "'Local' law enforcement isn't that local anymore."

United States Representatives

• Rep. Keith Self (R-TX), Wylie's own congressman, introduced the PRIVACY Act to require a warrant before federal agencies can access ALPR data.
• Rep. Tim Burchett (R-TN) introduced the Protection Against Mass Surveillance Act, which would bar federal agencies from purchasing, operating, or even accessing ALPR systems.
• Rep. Eli Crane (R-AZ) and Rep. Andrew Clyde (R-GA), cosponsors, called Flock part of "the federal government's dangerous web of warrantless surveillance."
• Rep. Shontel Brown (D-OH) joined Sen. Wyden in asking the Government Accountability Office to investigate failures to protect surveillance data.

Texas Legislature

• State Rep. Brian Harrison (R-TX), who voted for the fee that ended up funding Flock, said he "never could have even contemplated that this would be used to fund what is effectively warrantless surveillance."

Interstate 94, Wisconsin - August 2026

Amber Newell was surrounded by Brookfield police on I-94, held at gunpoint and handcuffed on the interstate while bystanders filmed, because a Flock alert flagged her car as wanted. The alert was stale: Milwaukee PD had closed the investigation but never removed her vehicle from the system. It was the second time in one week she'd been stopped at gunpoint on the same dead alert; the first time, police towed her car and released her without ever explaining why.

Brookfield's police chief said his officers acted "consistent" with the information available. That's exactly the problem: officers treat every alert as real, so when the database is wrong, an innocent driver meets drawn guns.

Source: FOX6 Milwaukee, August 2026

Aurora, Colorado - 2020

A mother and her 6-year-old daughter were pulled over at gunpoint and forced to lie on hot pavement. Police had guns drawn on a kindergartener because an ALPR system confused their Colorado car license plate with a stolen Montana motorcycle.

The city paid a $1.9 million settlement in 2024 for the trauma inflicted on this family.

Española, New Mexico

A 12-year-old was handcuffed when the ALPR system misread a license plate digit, confusing a "2" for a "7."

One month later in the same town, a 17-year-old honors student was detained at gunpoint based on another mistaken vehicle match.

Oakland, California - 2018

Brian Hofer was held at gunpoint after his vehicle was incorrectly flagged as stolen by an ALPR system. He wasn't a criminal, just a privacy advocate who had testified against surveillance systems.

Flock Isn't a Camera Company. It's a Surveillance Platform.

The phrase "license plate reader" makes this sound like a single gadget. It isn't. The camera is the entry point to FlockOS, a platform built to pull every surveillance feed in a jurisdiction onto one live map. Flock's own developer documentation describes a "vendor-agnostic" system designed to tie together "vehicles captured by ALPR, video cameras, patrol vehicles, body-worn cameras, and more," plus third-party alarms, drone-detection systems, GPS trackers, and 911 computer-aided dispatch.

Flock's product line now spans license plate readers, audio and gunshot detection, live pan-tilt-zoom video (Condor), drones marketed as "Drone as First Responder," a people-search product (Nova) that links plates to identities, and this integration layer that ingests outside tracking feeds. New capabilities arrive by software update: no new camera on a pole, and no new council vote.

And the platform just grew again, six days before Wylie's vote. On August 19, 2026, WIRED revealed Flock is testing an AI tool called OS Investigate (internal name: Nightshift) that lets officers search for people by how they drive: no license plate, no name, no crime required. Through a chatbot interface, it can flag "potential witnesses" by how often a car passes through a neighborhood, surface a driver's "associates" from the cameras they pass together, and pull in commercial identity databases. For years, Flock told the public its cameras "cannot recognize, identify, or track individuals." This tool is built to do both, and reporters found it in files exposed on Flock's own login pages.

Why this matters for the decision in front of Wylie: contracting with Flock is not buying a camera. It is subscribing to a surveillance-platform company whose capabilities expand on its own timeline. A city can reasonably decide what surveillance it wants, but it should decide that with eyes open about what the vendor actually is, and it should decide it in public. That is the entire reason to end this contract and start any future conversation with a bid, a hearing, and a warrant requirement.

Sources: Flock Safety Developer Platform documentation (Inbound Alerts API; Platform Concepts), captured and archived August 11, 2026; Engadget on WIRED's OS Investigate investigation (Aug 2026); Gizmodo. Quotations describe Flock's product platform, not the specific equipment deployed in Wylie.

We Read Flock's Own Crime Study, Too

When Flock and city officials say the cameras "solve crime," they are almost always pointing at one document: Flock's 2024 white paper, "An Initial Evaluation of Effectiveness in Aiding Police in Real-World Crime Clearance." It's the source of the "9.1% increase in clearance rates" number you'll hear at council meetings. We read all 21 pages. Here's what the marketing leaves out.

It was written by Flock, for Flock. Both authors are Flock Safety employees. It was self-published as a "white paper," not peer-reviewed, and the underlying data can't be independently checked because it was collected on the condition that it only ever be shown in aggregate.

The data is Flock's customers grading their own homework. It's a survey of Flock customers, asked "How many arrests have you made that can directly be attributed to the use of a Flock Safety camera?": self-reported, self-attributed, from the agencies most invested in the answer. Of 477 contacted, only 123 data points survived into the model.

And the study itself admits it cannot prove the cameras solved anything. In its own words:

"…it must be established that these are crimes that otherwise would have remained unsolved in the first place. Given the state of data collection on this issue, it is unlikely to be solved at scale via quantitative analysis."

That is the whole question. Would these crimes have been solved anyway, by ordinary police work? Flock's own study concedes it hasn't answered that, and probably can't with numbers. The paper calls itself "preliminary," "exploratory," and says its findings only "suggest a positive impact." Even the celebrated 9.1% comes with fine print the marketing drops: it's per additional camera for every single officer, and it measures the "ALPR-assisted clearance rate" (the share of cases the agency itself credited to a camera), not a drop in crime.

So the strongest evidence for Flock, produced by Flock, does not claim the cameras reduce crime, and admits it can't show they solve crimes that wouldn't have been solved otherwise. The independent study that did test it (Atlantic City, above) found no drop in violent crime. Read Flock's paper yourself:

Source: Flock Safety White Paper, 2024 (full PDF, archived). Snow & Charpentier, Flock Safety. Quotations verbatim from the paper's abstract, Data Validation, Results, and "Avenues for further research" sections.

Even the outside academic Flock leaned on to lend the study credibility has since said the data wasn't good enough to use. The white paper credits two outside consultants for lending it academic weight: Johnny Nhan, Ph.D. (Texas Christian University) and Richard C. Helfers, Ph.D. (UT Tyler). 404 Media investigated and found Nhan now disowns the result: "I personally would have done things much differently," he told reporters, saying the police-collected data was "too varied and incomplete for us to do any type of meaningful statistical analysis." Records 404 Media obtained show Flock steered which departments Nhan contacted, and an email in which Nhan asked a department for data showing a "big swing" in clearance rates before and after adopting Flock, sourcing that looks results-oriented, not neutral. Nhan has since stopped doing further research for Flock.

And there's now a real-world test of the paper's central claim (more Flock cameras should mean higher clearance rates) in the single largest Flock deployment in the country. Atlanta's camera network grew from 3,300 to 28,626 cameras between 2021 and 2026, per the Atlanta Community Press Collective's analysis. The same analysis, built on FBI-verified crime data, found clearance rates across most major crime categories stayed flat or fell over that same span: homicide clearance dropped from 53.4% to 48.0%. Atlanta's police chief separately claimed a 77.55% clearance rate for 2025; the FBI's own numbers for the same year show 48%.

Additional sources: 404 Media, "Researcher Who Oversaw Flock Surveillance Study Now Has Concerns About It"; Atlanta Community Press Collective, July 28, 2026

There Was No Crime Wave to Stop

Cameras like these get sold as an answer to rising crime. So it's worth asking what crime was doing in Wylie when the contract was signed on August 13, 2024. The answer, using only the city's own numbers: falling. Wylie had just been named the 12th safest city in Texas, its best ranking ever, "up 6 spots from 2023" in the city's own words. Months later, Wylie PD's annual report confirmed 2024 crime per 1,000 residents was down 4% and calls for service were down almost 10%. That report doesn't mention Flock once, because no camera had been installed yet.

And what did the city say the cameras were for? The purchase records answer that too. The $32,800 buy was a line item inside a routine budget amendment on the May 28, 2024 consent agenda, paid from seized funds. The entire written justification in the August 13 packet is one paragraph calling the cameras "a valuable tool in identifying vehicles involved in various criminal activity." No crime statistic appears anywhere in either document. Not one number, before or after. A city at its safest point on record bought a mass surveillance system without ever writing down what problem it was solving.

Two years later, there is still no data. The first cameras went live in late 2025, so no annual before-and-after comparison even exists yet, and the city has never published one. If the cameras were the reason Wylie is safe, someone would have to explain how the city earned a decade of safest-city rankings without them.

Sources: City of Wylie, Recognitions & Awards (SafeWise rankings as claimed by the city itself); Wylie News, "Wylie police report 4% decrease in 2024 crime" (Feb 13, 2025); 05-28-2024 City Council Agenda Packet, Item H (Ordinance 2024-16); 08-13-2024 City Council Agenda Packet, Item B (obtained by public records request, on file).

Times Flock's own status page admits the "hotlist" alert pipeline was delayed or down, in 18 months. The real-time alert is the entire sales pitch: the stolen car, the AMBER Alert, the recovery "in under ten minutes." Flock's public incident log (50 incidents from March 2025 through August 2026) shows that exact feature failing about once a month, including an April 2026 pause of hits from the FBI's NCIC database, and two 2026 degradations of the audit and transparency portals cities rely on for oversight.

Notice what never goes down: the scanning. The collection is reliable. The safety features are not.

Source: status.flocksafety.com incident history, captured August 11, 2026 (archived). Details on the Fine Print page.

Potential misread records created every month, using Flock's own numbers. Flock advertises 96% character-level read accuracy "under optimal conditions." Flock's network runs roughly 20 billion scans a month. Take the company entirely at its word and do the arithmetic: 4% of 20 billion is up to 800 million wrong records a month, wrong plates attributed to real cars, sitting in a database police search as truth. And that is the conservative reading: 96% is character-level accuracy, and a plate has about seven characters, so the share of whole plates read perfectly is necessarily lower still.

And "optimal conditions" is the best case. When Roseville, California audited its actual Flock cameras over two years, 71% of reads were wrong. Flock blamed camera placement and aging hardware. That's the point: the advertised accuracy depends on deployment conditions nobody audits. A misread isn't a harmless glitch. A misread is how a Colorado mom's sedan became a stolen Montana motorcycle, and how Amber Newell ended up handcuffed on I-94.

Sources: Flock Safety's published accuracy specifications, as reported in "Flock Says Its Cameras Are 96% Accurate — Police Found 71% of Alerts Were Wrong"; Roseville, CA field-accuracy analysis; ~20 billion scans/month network volume

The "100% Homicide Clearance Rate" Story Flock Tells Every City, Fact-Checked

Flock's sales materials repeatedly point to two cities that hit a 100% homicide clearance rate (Tulsa, OK and Cobb County, GA) as proof the cameras work. Both numbers are real. Neither holds up as evidence Flock caused them.

Tulsa cleared all 50 homicides it investigated in 2024, and the department's own announcement confirms it. But independent investigative reporting found Tulsa's homicide clearance rate had "consistently exceeded 90%" for the entire decade before any Flock camera was installed there in 2022. Tulsa hit 100% once before, in 2018, four years pre-Flock. A Duke University criminologist quoted in the piece attributes the city's persistently high rate to its homicides mostly being domestic disputes with a known suspect, not to any technology. Tulsa PD's own official statement credits "community support, new technology, and the hard work of detectives." It does not name Flock. The rate fell back to 94% the very next year.

Cobb County's 62-for-62 homicide record (March 2021 into early 2023) is also genuine, and traces to the department itself, not Flock. But the police chief's own quote to local press credited the result to a combination of tools (real-time crime center cameras, facial recognition, license plate readers) plus, in his words, "detectives who sacrificed hours and hours... and exhausted all leads without fail." Flock's marketing compresses that multi-factor, department-credited outcome into a single-technology success story implying its cameras were the decisive factor.

The best available check on self-reported "we solve everything" clearance claims comes from the same metro Cobb sits in: Atlanta, where the police chief's claimed 2025 homicide clearance rate ran nearly 30 points above the FBI's verified figure (see the crime-study section above).

Sources: Tulsa Police Department; The Frontier, "Tulsa's Surveillance Gamble"; KTUL (2025 rate: 94%); FOX 5 Atlanta (Cobb County); Atlanta Community Press Collective (77.55% claimed vs. 48% FBI-verified)

The Same Anecdote, Copy-Pasted Onto Companies With No Confirmed Flock Relationship

Flock builds individualized sales landing pages for target companies, each showing "results" as if that company already uses Flock. Checked against Flock's own official customer list, several of these named targets, including Nike and Disney, appear nowhere on it, and no independent reporting confirms either company has ever had a relationship with Flock. The pages read as cold outreach ("Talk to Sales," "mitigate PR risks for Disney"), not case studies.

What sells the pitch is one recycled statistic: a "100% reduction in parking lot crime," lifted from a single real case (a Memphis cancer center that installed three cameras, no stated timeframe, one hit-and-run solved) and pasted as a generic stat tile onto unrelated pages built for Disney, an electric utility, and a paper company. A second stat, a "97% success rate" at an unnamed "Alabama retailer" (a single undisclosed storefront, one 90-day quarter, three theft cases solved), is copy-pasted across personalized pages for Nike, Walmart, Burlington, and Publix, relabeled inconsistently between copies.

If a Wylie council presentation or a Flock sales rep cites a specific "case study," the pattern above is worth remembering: the same handful of small, short-window anecdotes get reused everywhere.

Sources: Flock Safety official customer list; West Cancer Center case study; "Alabama retailer" case study. The personalized sales pages for Nike, Disney, Walmart, Burlington, Publix, and others were captured and archived by this campaign on August 14, 2026, and remain available on request.

The Audit Log Says What the Police Won't

On February 10, 2026, Assistant Chief Walters told City Council that officers can't misuse the system, claiming Flock now requires officers to "select a crime plus a case number" and "can't just make up" a reason for searching.

One month of Wylie PD's own audit logs (January 7 – February 6, 2026) tells a different story. Justifications officers actually entered include:

• The case number entered literally as "TEST" on a "Robbery - follow up" search, in a live system tracking real Wylie residents
• "kuria", an unexplained entry appended to the justification on more than one search (for example, "Drugs/Narcotics - kuria")
• "case is from 2021, looking for recent hits", a five-year-old case triggering a nationwide camera search
• Generic, untraceable reasons like "investigation", "suspect", "theft", with no further detail

Source: Wylie PD Audit Log, FOIA release PIR-2026-78, covering 1/7/2026 – 2/6/2026.

The number of cameras a single Wylie officer's "hit and run" search reached. One query, 6,119 networks, 90,911 devices nationwide. This was not an unusual entry. Multiple Wylie searches in the audit log scanned tens of thousands of cameras across the country for routine local cases.

The Council was told this is a tool for "vehicles entering and exiting our city." The audit log shows it's being used to scan a national surveillance network on the basis of a single Wylie file number. And the network runs both ways: when Lucas County, Ohio audited who was searching its cameras, it found Houston PD alone had run more than 71,000 searches against the county's data in a single month, with no clear explanation. (13abc, August 22, 2026)

We Read Flock's Fine Print. You Should Too.

Flock's public FAQ promises it will "never share or sell" your data, and won't even "access" it. Flock's actual contract takes a "limited, non-exclusive, royalty-free, irrevocable, perpetual, worldwide license" to use customer data for Flock's own product development. Its LPR policy reserves the right to hand data to "government officials and/or third parties" on Flock's own "good faith belief." Its liability, if everything goes wrong, is capped near one year of fees (about $27,000) under Georgia law, in private arbitration. And its own policy tells officers that plate reads "may be incomplete or inaccurate."

We archived every document with cryptographic hashes on August 11, 2026, so the record survives the next revision. Read the side-by-side:

The Movement Has Reached DFW

Wylie isn't going first, and it isn't going alone. As of summer 2026, organized resident campaigns are active across the metroplex and Texas:

• DeFlock DFW: the metroplex-wide umbrella project uniting the city campaigns, documenting ALPR programs across Dallas, Fort Worth, Denton, Frisco, Carrollton, and McKinney. Holds regular community meetings across the metroplex
• DeFlock Carrollton: 600+ petition signatures demanding removal of the city's roughly 70–95 Flock cameras (about $261,000/year)
• Richardson: resident petition launched July 2026 demanding removal of existing units and a public Fourth Amendment commitment from the mayor and council
• DeFlock Tyler: countywide petition against all ALPR vendors in Smith County, with public info meetings underway
• El Movimiento DFW: running know-your-rights workshops on ALPRs and immigration enforcement across Dallas–Fort Worth
• Texarkana, College Station, and Harris County: residents packing council and commissioners meetings to demand cancellation

Nationally, DeFlock.me now maps roughly half of Flock's ~100,000-camera network, and an estimated 30 local DeFlock-style groups exist across the country. The National Week of Action Against ALPRs runs August 16–22, 2026, with residents in 75+ cities showing up to council meetings, canvassing neighborhoods, and demanding contract cancellations.

Sources: KERA News (Carrollton); Dallas Observer (DFW overview); KLTV (Tyler); KBTX (College Station)