The Flock is breaking up, and America finally started asking who was watching. In Part 1, when we first talked about Flock Safety Cameras, I wasn’t trying to convince you that every little black camera sitting beside the road was secretly plotting the downfall of civilization. Flock cameras do some good things. That’s precisely why law-enforcement agencies bought them. They’ve helped recover stolen vehicles, located missing people, identified suspects, and solve violent crimes. They can take something that might have required hours – or days – of traditional police work and reduce it to minutes. If somebody kidnapped my child and police could locate the vehicle through Flock, I promise you I wouldn’t be standing at the police station demanding a philosophical discussion about data-retention policy. And THAT’S precisely what makes the conversation difficult.
Bad technology is easy to regulate…
But powerful, useful technology is much harder. Because now we’re trying to preserve something genuinely valuable… while simultaneously deciding how much power we’re comfortable attaching to it. And before we even published the first article, something interesting was happening. The average citizen began asking about these cameras. It was the kind of questions Flock’s executives tried to head-off with reassurances… that were really false narratives. They gave answers that didn’t satisfy, and even some cities didn’t like those answers.
According to data compiled by surveillance watchdog Secure Justice and reported by Ars Technica, 214 cities & counties had dropped Flock since 2021. With 165 of those cancellations occurring in 2026… and 90 happening during August alone. That’s roughly 3 communities per day dropping Flock during the month of August 2026.
But before anybody starts writing Flock’s obituary, don’t. Flock says new municipal partnerships still substantially outpace cancellations. The company remains enormous, its cameras remain all over America, and plenty of law-enforcement agencies remain enthusiastic about the technology. Flock isn’t collapsing. The honeymoon is ending. And those are two very different things.
The Flock Camera Was Never Really the Problem…
This is where I think we need to correct something in the entire Flock debate, including some of the ways I’ve discussed it myself. We’ve spent too much time talking about cameras. A camera isn’t particularly frightening. I’m driving through town in my F-150 and a police officer sitting at an intersection sees me: Ruby-red Ford, lifted, big tires, plate number, driver probably listening to something louder than necessary. The officer watches me drive away. Nothing particularly interesting happened.
A Flock camera sees essentially the same thing. And again, nothing revolutionary. Except the problem begins when that observation doesn’t disappear. It’s recorded, stored, and made searchable. Then another camera sees me. Then another. Another town. Another county. Potentially another state. Eventually – the question stops being, “Did this truck pass this camera?” and becomes, “Where has this truck been?” Those are not the same question.
And once thousands-upon-thousands of Flock cameras can communicate through enormous networks of agencies…
We aren’t really discussing cameras anymore. We’re now discussing infrastructure. Welcome to Alpharetta, Georgia. If you want to understand what I’m talking about, forget Washington, D.C. for a minute, and just come to Georgia. Alpharetta is exactly the kind of place where Flock makes sense. It’s an affluent Atlanta suburb with busy roads, neighborhoods, shopping centers, vehicle theft, and major transportation corridors nearby. You can practically write the sales pitch yourself: Put cameras around town, someone steals an Escalade, Escalade passes camera, police locate Escalade, bad guy’s busted. And everyone (except the bad guy) goes home happy.
Here the problem though, records obtained by WIRED showed that Alpharetta’s Flock information has become accessible to more than 2,000 organizations across the United States… while Alpharetta could receive information from more than 1,300 organizations in return. And we’re not talking about Alpharetta just sharing with neighboring towns/counties. We’re not even talking about every police department in Georgia. The sharing ecosystem reportedly included police departments, universities, airports, government investigators, and task forces scattered all around the country. At some point, if I’m an Alpharetta resident, I think I’m entitled to ask a pretty simple question:
Why does the entire country potentially need access to information about my truck driving through Alpharetta?
There are perfectly legitimate reasons for law-enforcement agencies to share information. Criminals do not respect municipal boundaries. A stolen vehicle doesn’t reach the Fulton County line, pull over, and say, “Damn jurisdiction!. Guess we’re done boys.” Kidnappers cross county lines, car thieves cross state lines and organized theft rings operate across the country.
That interconnected capability is exactly what makes Flock extraordinarily powerful… it’s also exactly what makes Flock extraordinarily complicated.
Alpharetta Police told WIRED that its settings automatically approved sharing requests from law-enforcement agencies within a 500-mile radius. Five hundred miles. At that point I’m not sure “neighboring agencies” is the phrase we’re looking for anymore. That’s more like, “Would you like the Southeast with that?”
To be fair to Alpharetta PD – and fairness matters throughout this article – the department said it wasn’t sharing live feeds or analytics with those agencies. But the larger question remains: How many people involved in approving these cameras originally understood what the system could eventually become?
Because when communities first entered this world, the concept sounded considerably smaller. A camera sees a car. Crime happens. Police investigate. Police check camera. Simple.
Now fast-forward several years – and thousands of organizations can potentially be connected through these networks. Nobody held a ribbon-cutting ceremony. And nobody stood behind a podium and announced, “Ladies and gentlemen, today we begin construction of a nationwide vehicle-location network.” It just… grew. One camera became another, one department connected to another, etc. And eventually, after years of adding individual pieces that each seemed relatively small on their own, somebody finally zoomed-out and realized just how massive the whole thing had become.
And That’s Where Things Get Weird…
While Alpharetta was connected to this enormous network, investigators say one of its own officers was using Flock for something considerably more personal. Former Alpharetta officer Dustin Bozzo allegedly searched for the vehicle belonging to a former romantic partner – who was also an Alpharetta police officer – 56 times between March and May 2026. Then he apparently noticed another vehicle around hers. That vehicle belonged to another officer, so investigators say he searched that vehicle too – 29 times. After his former partner was involved in an accident and began borrowing her uncle’s vehicle, investigators say that vehicle was searched another 18 times over just 9 days.
Some of those searches were reportedly entered under reasons including “Wanted Person” and “Traffic Infraction.” The problem is, investigators ultimately concluded there wasn’t a legitimate law-enforcement basis for those searches.
Bozzo reportedly told investigators he wasn’t trying to intimidate or harm anyone, and that some searches were motivated by wanting to know his former partner arrived home safely. That’s important context, but strangely, I think it makes the example more useful. Because when people imagine abuse of surveillance systems, they usually imagine some mustache-twirling villain sitting in a dark government bunker surrounded by monitors. Real life is usually considerably less cinematic. Sometimes it’s jealousy, curiosity, concern, a breakup, anger, an employee having a terrible day. One bad decision… and then another… but now with access to an incredibly powerful database.
Alpharetta placed Bozzo on administrative leave. He later resigned, and the department opened a criminal investigation. That’s what accountability is supposed to look like.
But then I found the part of this story that bothered me considerably more…
According to Alpharetta Police, Chief Trent Lindgren learned about Flock’s Audit Assistance feature in July. He activated it, and by July 20th, the department had opened its investigation into Bozzo. Audit Assistance is designed to identify abnormal behavior within Flock usage. Flock has since announced that the feature will become mandatory, along with additional safeguards involving authentication, required case codes, proactive account lockouts and shorter recommended/default retention.
Excellent. Those are good changes. I’m not going to criticize a company for implementing protections I’ve been arguing should exist. But I do have one rather obvious question: Why wasn’t abuse detection mandatory from the beginning?
Reporting on Flock’s changes indicated only around 1/3 of customers had previously activated Audit Assistance. We’re talking about software capable of searching vehicle movements across enormous geographic areas, and the system designed to notice unusual searches was optional. That’s like selling a chainsaw and putting the chain brake in the Technology Package. Alpharetta accidentally demonstrated why the feature matters. Turn auditing on, suspicious behavior gets identified, an investigation begins, and an officer resigns. That’s not evidence that the safeguard failed. That’s evidence that the safeguard should have been there the whole time.
The Flock Camera Is the Eyeball…
Alpharetta is not an isolated example of just how large these sharing networks can become. WIRED also examined Reynoldsburg, Ohio, where police leadership reportedly believed officers could search cameras only within Ohio, only to discover that Nationwide Lookup functionality had been active.
When Reynoldsburg asked Flock when nationwide access had been enabled, the precise date reportedly couldn’t be determined from older logs. Flock explained that Nationwide Lookup had once been a common default configuration. Reynoldsburg subsequently restricted searches to Ohio.
Then there was an audit record showing one Reynoldsburg officer conducting a search across 53 Flock networks. Roughly 40 minutes later, a search using the same offense code reportedly went across 5,392 networks. At that point calling Flock a “license plate camera” feels a bit like calling The Internet a telephone accessory. Technically, somewhere in there, you’re not entirely wrong. You’re just missing virtually everything important.
That little black box beside the road isn’t the impressive part. Honestly, most Flock cameras look like somebody zip-tied a solar-powered birdhouse to a pole. The impressive part is what you can’t see: The database, permissions, searches, audit logs, sharing relationships, agencies, etc. And the ability for somebody potentially hundreds of miles away to ask whether a particular vehicle passed one of those little birdhouses.
The Flock camera is the eyeball; the network is the brain…
Current reporting puts Flock’s network at roughly 120,000 cameras across 49 states, processing an enormous volume of vehicle observations every month. The overwhelming majority involve people who haven’t committed a crime. They’re going to work, picking up children, going to Walmart, dinner, driving home. You know… just living.
That doesn’t automatically make the system unconstitutional. It doesn’t automatically make it malicious. But it should absolutely make us interested in who can search it.
“But I’ve Got Nothing to Hide” Here we go. My favorite surveillance argument. 99% of the populous has nothing to hide and neither do I. But I still close the bathroom door, and there isn’t a meth lab in there. I put passwords on my accounts, and I’m not coordinating an international jewel heist. My wife and I occasionally have conversations that aren’t anybody else’s business. That’s called being human. Wanting privacy doesn’t mean you’re hiding criminal activity. It means you want the ability to decide which parts of your ordinary life belong to everybody else.
This is where people get hung up on Flock. “But you’re driving on a public road.” Correct. A police officer can see me. You can see me. Some guy walking his dog can see me. That’s not the issue. The issue is persistence. One person seeing my truck isn’t particularly revealing. Hundreds of observations assembled chronologically can be. Scale changes things. Searchability changes things. Aggregation changes things. And eventually – enough ordinary observations can begin producing an extraordinary amount of information.
Is Flock Unconstitutional?
This is where we need to be careful, because I don’t want this becoming Facebook Law School. A federal court has already rejected one major 4th Amendment challenge against Flock. In Schmidt v. City of Norfolk, Virginia residents challenged Norfolk’s ALPR network as warrantless mass surveillance. In January 2026, the federal district court granted summary judgment to Norfolk. So no, a court has not simply declared Flock cameras unconstitutional.
But that isn’t the end of the legal conversation either. The plaintiffs appealed, and the underlying facts explain why this issue isn’t going away. Court records indicated one plaintiff’s vehicle had been photographed hundreds of times over a period of several months. Again, one observation isn’t particularly remarkable, but hundreds? Now we’re having a different conversation.
American courts have spent years wrestling with the reality that modern technology can transform individually mundane pieces of information into something enormously revealing when collected at scale. ALPR networks are going to force courts to continue defining where that line exists. And as usual, technology got there before the law did. Flock built the thing. Now everybody gets to figure out the rules. Fantastic planning. Feels real dystopian.
“What If Somebody Abuses Flock?”
This may be the biggest change since Part 1. For years, critics of large surveillance databases asked, “What happens if somebody abuses it?” And supporters could reasonably respond, “Do you have evidence anyone actually will?”
Well… we can retire that argument. A Washington Post investigation identified dozens of law-enforcement officers who HAD been charged with (or accused of) unauthorized license-plate-reader use, including allegations involving officers searching for women or romantic partners. We’ve already discussed Alpharetta, and we don’t need to turn this into The Encyclopedia of People Who Shouldn’t Have Touched the Search Bar. But Alpharetta isn’t alone. Cases or investigations have emerged in Arizona, Texas, Florida, Massachusetts and Wisconsin – involving alleged personal or unauthorized searches.
Then there’s Lufkin, Texas, where former officer Zachary Anthony Klein was indicted in August on 100 counts of misuse of official information involving alleged searches connected to 11 people. Those remain allegations, and accused people are entitled to due process. The public-policy question, however, no longer requires speculation. Can powerful law-enforcement databases be used for personal reasons? The answer is yes. So the conversation has to evolve.
How difficult can we make Flock misuse, and how quickly can we detect it?
CEO Garrett Langley has acknowledged that stronger safeguards are necessary, and Flock has announced significant changes. Its recommended/default ALPR retention period is moving from 30 days to 7. Audit Assistance is becoming mandatory. Case numbers or codes will be required for searches. Authentication is being strengthened. And suspicious activity can trigger lockouts.
Good. Full stop. If the purpose of criticism is to improve something, then you don’t get to complain when the company improves it. But – improvement doesn’t automatically erase the questions that produced it. Why wasn’t abnormal-use detection mandatory previously? How did some agencies apparently not understand the geographic reach of their search capability? And why were nationwide searches ever a common default configuration? How easily can municipalities determine every organization receiving their information? Reducing retention from 30 days to 7 helps, however But it doesn’t answer everything else.
Trust Is Not a Security Feature…
“I trust our police department.” Great. But that’s irrelevant. I trust my friends. But they don’t have administrator access to my bank account. I trust employees. But they still have passwords & permissions. I trust pilots. Yet airliners still have redundant systems. Security isn’t an accusation.
