The Wild West of Roadside Cameras: Industry Insider Weighs In on License Plate Readers, AI, & Public Oversight
The cameras are only part of the story. An exclusive interview examines the monitoring, data collection, and oversight questions surrounding America's expanding roadside camera networks.
Flock Safety has become the public face of America's growing debate over automated license plate readers. But according to an engineer who has spent more than a decade building roadside camera systems, the national conversation is focused on only one piece of a much larger surveillance network.
“I don’t want to live in a world in which your every move is monitored and you have no idea where the data is going or what’s going on.”
Jen's Two Cents is the first news outlet to receive advance access to the white paper of an industry veteran writing under the pseudonym Edward Telleed, as well as the first exclusive interview about the report.
The White Paper | The Wild West of Roadside Cameras: An Insider’s View of Uninformed Consent and Public Surveillance
Download the report for free here.
Throughout our interview, Telleed argues the technology itself is not the central issue. Instead, he says the public is only beginning to understand how widespread roadside camera networks have become, how artificial intelligence is changing their capabilities, and why he believes greater transparency and public oversight are needed.
“It isn’t until now with the Trump administration and ICE and Flock that everybody’s paying attention. So this is why I’ve decided to do this.”
Beyond Flock: A Growing Network of Roadside Cameras
Flock Safety may be drawing the headlines, but it is only one of several companies manufacturing automated license plate reader systems deployed across the United States.
Roadside cameras have been installed for years on highways, intersections, neighborhoods, parking garages, and commercial properties. As the technology becomes more visible, public attention has shifted toward questions surrounding how these systems operate, what information they collect, who can access the data, and what role local communities play in approving their deployment.
Unlike privacy advocates or government officials commenting from the outside, Telleed approaches the issue from inside the industry.
After working in the security camera business, he writes that thefts at his family’s business in 2012 prompted him to begin developing roadside camera technology. His first roadside camera was deployed in 2013, followed by experiments with automated vehicle identification the following year. In our interview, Telleed said he is based in the northeastern United States and that his company now installs roadside camera systems in three states, primarily serving private businesses while occasionally providing law enforcement access to privately owned camera footage during investigations.
When I first started in 2013 the accuracy of plate readers was around 80%. Now in 2025 it is 95%. This is due to improvements in algorithm and improvements in camera. The take away here is that this is not that it is as good as facial recognition, but your license plate, unlike your image…is always guaranteed to map to you and identify you.”
The Debate, He Says, Isn’t the Cameras
Although Telleed helped build the technology, he says his concerns center on how roadside camera systems have been introduced into communities.
He says his company has historically sought public input before installing publicly visible systems… even when that meant projects were rejected.
“Roadside cameras work, and I’m a proponent of them for what they’re supposed to be doing,” Telleed said. “However, the problem is that over the past several years, nobody has been asking permission. They haven’t been going through ethical channels. Just because you don’t have to ask doesn’t mean you shouldn’t.”
He says communities should have the opportunity to discuss surveillance technology before it is deployed.
“Sometimes we’ve gotten rejected. But that’s just the way it is,” he said. “These things are going in wholesale now. Very quietly. Now all of a sudden they’re coming to light and we have these things. It’s not the way to do it.”
As artificial intelligence continues to evolve, Telleed believes public involvement becomes increasingly important.
“With the power of AI and everything that’s advancing, we’re gonna run into real trouble unless we have public input, public discourse and some control, regulation and input on these things.”
Why Location May Matter More Than Identity
Facial recognition often dominates public discussion surrounding surveillance technology.
Telleed argues that, at least today, that focus may overlook what roadside camera systems are designed to do best.
He says automated license plate readers are optimized to identify vehicles and document where they travel. According to Telleed, repeated observations of a vehicle can reveal behavioral patterns over time without ever identifying the driver.
“People should not be worried, at least at this point in the game, about facial recognition more than the pictures of the car and the license plate,” Telleed said. “Your license plate and your car are the low-hanging fruit to the data. Your face can be anywhere. But I’m like ‘who are you’? I just run your license plate and I’ve got everything about you.”
To illustrate the distinction, he offered an analogy.
“Everybody knew what Osama bin Laden looked like. That wasn’t the problem. It was finding out where he was. Where you are is more important than who you are.”
In his white paper, Telleed argues that the real power of roadside camera networks lies not in a single image, but in the accumulation of data over time. Drawing on the legal concept of Mosaic Theory, he writes that repeated observations from multiple cameras can establish a person's "pattern of life," allowing analysts to infer routines, destinations, relationships, and other behavioral patterns from otherwise ordinary travel data. (Section 2.1, p. 18.)
“When you have multiple cameras and multiple frames of reference, you can deduce a tremendous amount of information from a photograph,” he said.
He argues that approximately thirty days of vehicle location history can establish routines, travel habits, and other recurring behavioral patterns.
Before automated license plate readers became widespread, tracking a vehicle often required a court-authorized GPS tracker or physical surveillance by investigators.
Today, Telleed argues that automated camera networks have fundamentally changed that equation.
“With the Flock system—and all technology, all of the companies, Axon, Motorola, everything else now—you can do all that to every single car in America every second for pennies on the dollar.”
Why Facial Recognition May Be the Wrong Question
Drawing on his experience designing roadside camera systems, Telleed says camera placement, not simply camera quality, is one of the biggest limitations on facial recognition.
“One of the major limitations of any camera is its position/location.” he writes. “It is possible a license plate reading camera could see a face through a front window of a vehicle, but it is generally not positioned to do so. Its angle, field of vision and overall position and location is to capture images of vehicle and their plates.”
During our interview, Telleed said many modern camera platforms contain processors capable of running facial recognition and other AI applications. However, he says those capabilities depend on collecting usable images.
“The camera has to be positioned correctly. If it’s not, you’re not going to get reliable facial recognition.”
Instead, he argues that license plates often provide investigators with more immediately actionable information than facial images.
The Next Phase of Roadside Surveillance
Telleed argues that the capabilities of roadside camera systems are expanding rapidly through what he describes as three forms of escalation.
The first is technological: cameras continue to improve while artificial intelligence becomes more capable and equipment grows less expensive.
“The cameras get better. AI gets better. Cameras get cheaper, smaller, and more capable.”
The second is political, as more agencies adopt technologies that simplify investigations.
The third is criminal, as offenders adapt to increasingly sophisticated surveillance systems.
“It’s not really stoppable,” he said. “Technology offers the easiest path to crime prevention, revenue generation, everything. Policing. It’s the easiest thing to do and it’s the path of least resistance.”
Who Watches the Watchers?
Rather than calling for an end to roadside camera technology, Telleed says the greater priority should be transparency and oversight.
One area he identifies is public access to information.
“The more immediate concern—you have to make sure that we don’t cripple the Freedom of Information Act too much. You have to have transparency... the ability to watch, monitor, and supervise these data centers.”
Because surveillance laws are primarily governed by individual states, he encourages citizens to pay close attention to legislation involving public records, audits, data retention, and oversight.
He also argues that communities should be notified whenever new surveillance capabilities are added to existing systems.
What Telleed Says Should Change
Looking ahead, Telleed advocates what he describes as a balanced approach. An approach that allows the technology to continue evolving while establishing clearer safeguards for its use.
Section 2 of Telleed's white paper examines the legal framework surrounding automated license plate readers, outlining the types of legislation he believes should be adopted, as well as proposals he argues should be avoided.
He says public surveillance systems should be discussed in open meetings before deployment, and residents should be notified whenever significant new features are introduced.
“Up until this point, it’s been all about trust,” he said. “Those days are over.”
He believes future legislation should include transparency requirements, independent oversight, accountability measures, and penalties for misuse.
“You have to speak up now before this creeping functionality comes in and it becomes accepted and more powerful as it goes. Then there’s no turning back.”
He ended the interview with the same warning that opens his white paper:
“I don’t want to live in a world in which your every move is monitored and you have no idea where the data is going or what’s going on.”
Opportunity to Respond
The views expressed in this interview are those of the guest and are presented to encourage discussion on a matter of public interest. They do not necessarily reflect my views.
Representatives of Flock Safety, other license plate reader manufacturers, law enforcement agencies, or any organization discussed in this interview are welcome to contact me if they would like to respond or provide additional context in a future article or interview.
Flock CEO Acknowledges Past Remarks, Promises Greater Transparency
As public scrutiny of automated license plate readers continues to grow, Flock Safety CEO Garrett Langley says the company plans to engage more directly with critics and answer questions about its technology.
In a post published on X, Langley acknowledged that he had previously referred to members of the DeFlock movement as “terrorists,” calling the remark a mistake.
The CEO addressed DeFlock, describing it as an open-source project that has gained popularity in part because of what he characterized as Flock's inadequate response to public concerns surrounding automated license plate readers.
“The recent awareness surrounding their existence has also spiked a fair share of criticism. I think we have failed to answer these questions and concerns at scale, and we are going to do a better job of that moving forward.”
As part of that effort, Langley said Chief Strategy Officer Rahul Sidhu will begin addressing questions submitted by users on X, adding that both executives intend to engage more directly with the public.
Jen’s Two Cents.
I’m grateful to Edward Telleed for trusting me with the first interview about his white paper and for giving Jen’s Two Cents the first look at his research.
Whether you agree with his conclusions or not, one thing is becoming increasingly clear: the conversation around automated license plate readers is no longer just about one company. It’s about how governments, law enforcement, technology companies, and the public decide these systems should be used and what safeguards should exist.
Arizona lawmakers took a first step this year by introducing two bills addressing automated license plate readers. Neither made it to the governor’s desk. As legislators begin drafting bills for the next session, now is the time for the public to pay attention, ask questions, and contact their elected representatives.
Today’s Arizona primary election may change who is making those decisions. It doesn’t change the questions that still need answers.