When Tulsa police found a body dumped behind a car wash in 2022, there were no suspects and few leads. But within hours, investigators turned to a newly installed surveillance tool — one of the city’s first Flock Safety cameras — to identify a suspect’s vehicle. The car matched the description from a business security video and was the only vehicle of its kind picked up along the escape route. Officers traced the license plate, located the owner and got a confession.

That case, police say, was the first homicide solved using Tulsa’s now-ubiquitous Flock camera system. 

City officials called it one of the biggest advancements in Tulsa policing in decades. Tulsa has since invested millions of dollars in the artificial intelligence-enabled system, citing its role in solving homicides and other serious crimes. 

Today, the Tulsa Police Department operates 105 Flock Safety license plate readers and over 150 pan-tilt-zoom live-streaming cameras, managed from its real-time information center — a city-run surveillance hub that operates with Flock software. Over half of these cameras were installed in the past year. The company’s marketing materials showcase Tulsa’s surveillance network as a national success story and “a beacon for what modern, data-driven policing can achieve.”

Yet, nearly three years after its rollout, there is little evidence that the system has helped prevent and solve crime in Tulsa, according to crime data and interviews with experts. Recent academic studies suggest license plate readers offer few benefits, even for less serious crimes. And some researchers argue that funds may be better spent hiring more officers, particularly in short-staffed departments like Tulsa’s. 

The technology has also raised legal concerns: Tulsa installed Flock cameras despite a longtime state law that explicitly prohibits the use of license plate readers for any purpose other than the enforcement of the state’s mandatory insurance law. Civil liberties advocates argue that the cameras, which can track the movements of cars around the city, are a form of unconstitutional surveillance. 

A spokesperson for Flock Safety referred questions about the efficacy of license plate reader cameras in Tulsa to the Tulsa Police Department. 

Captain Jacob Johnston, who leads the city’s real-time information center, acknowledged that research has yet to show clear evidence the technology reduces crime.

“All I can say is, since we started putting this technology out, we’re seeing crime in the city of Tulsa go down,” he said. “We’re seeing benefits across the board and striving towards Mayor Nichols’ goal of making Tulsa the safest city in America.”

Since the cameras were installed in mid-2022, violent crime initially decreased but has since risen back to previous levels, according to Oklahoma State Bureau of Investigation crime statistics. 

Car theft and clearance rates have improved in recent years, but this trend began in 2021 — before the rollout of Flock cameras — following a post-pandemic surge in stolen vehicles and a subsequent crackdown, where police deployed bait cars, targeted repeat offenders, and increased patrols in car theft hotspots. Peer-reviewed studies from other cities have found license plate readers have no statistically significant effect on car theft clearances once confounding factors are taken into account.

Still, Johnston credits Flock Safety’s technology, which Tulsa was one of the first cities to adopt, with aiding in arrests across a range of offenses — from serious crimes like homicides and armed robberies to more common incidents like hit-and-runs and shoplifting. He said the system also helps officers quickly locate missing persons, such as people with dementia who wander from home. 

Johnston argues it has led to faster response times and more accurate identification of suspects, which isn’t represented in data on crime or clearance rates.

“If I have a homicide suspect that’s on the loose, the faster we can put handcuffs on them and keep the community safe, the better,” he said. “And so sure, I can see people saying we might have caught the person another way. But I don’t want to see them walking the streets any longer than possible.”

Nonetheless, a growing number of lawmakers, civil liberties advocates, and legal scholars argue that the benefits fail to outweigh the costs. They warn that these systems threaten foundational privacy rights and represent a worrying expansion of government surveillance.

“We are a nation of laws, not of technology,” said State Rep. Tom Gann, R-Inola, who led an interim study on license plater reader cameras in Oklahoma. “It just boggles the mind that you have chief law enforcement officers who are willing to defy the law because they feel like they have some sort of superiority, just because of what good they perceive they do.”

Last September, a McClain County judge ruled “unequivocally” against the use of license plate reader data as evidence in a criminal case, raising worries that their use may weaken or upend prosecutions in Oklahoma. 

The Tulsa Police Department did not respond to written questions about the legal concerns. Johnston said the Oklahoma statute that prohibits license plate readers was originally intended to regulate their use on turnpikes and argued that their broader deployment in cities has outpaced the law. He said he believes lawmakers should revisit the statute and noted that the department hasn’t faced legal challenges over its use of the technology.

Flock Safety’s cameras do more than read license plates. They also use artificial intelligence to create “vehicle fingerprints” based on make, model, color, and unique features like dents and bumper stickers — allowing the system to identify cars even when license plates aren’t visible. Officers can search for cars without a warrant using a license plate number or descriptors like “all vehicles with an American flag bumper sticker.” The software can then combine these datapoints to map a vehicle’s movements across the city.

On top of those they operate, ​​Tulsa police also have access to privately owned Flock cameras installed at shopping centers and in neighborhoods. 

When reporters visited Tulsa’s real-time information center in March, they gave Johnston one of their license plate numbers so he could demonstrate how the system works. Johnston typed the number into his phone and a detailed log of where the car had been over the past month came up in seconds. No warrant was needed. When the software prompted Johnston to give a reason for the search, he typed in “test.” 

Return on investment

City officials insist the system is critical for solving crime. 

“This is like turning the light switch on after our officers have been having to grope around in the dark for decades,” Mayor G.T. Bynum said in 2022. 

Since taking office late last year, Mayor Monroe Nichols has vowed to continue using data-driven policing to target crime hotspots in the city. Police Chief Dennis Larsen also plans to continue expanding the system, saying it has helped offset a staffing shortfall of about 140 officers. 

In January, Tulsa police credited Flock cameras with the department’s 100% homicide clearance rate in 2024, meaning every reported homicide led to an arrest.

“You wouldn’t be talking to me right now about a 100% clearance rate if we didn’t have the Flock cameras,” Lieutenant Brandon Watkins, Tulsa’s lead homicide detective, said in an interview with KRMG. Former police chief Wendell Franklin also said the cameras played a “huge role” in the department clearing 95% of its homicides in a Flock Safety advertisement in 2023. 

Flock Safety frequently sends press releases to Tulsa journalists crediting its technology for the city’s homicide clearance rate, and TV stations have aired stories based closely on these releases. 

However, FBI crime data shows that over the past decade, Tulsa’s homicide solve rate has consistently exceeded 90% — well above the national average — before the city installed Flock cameras in 2022. The rate was also 100% in 2018.

Duke University criminologist Philip Cook, an expert in crime clearance rates, attributes Tulsa’s high homicide solve rate partly to the nature of its killings, which often stem from domestic disputes rather than gang-related violence, making them easier to solve.

A spokesperson for Flock, Holly Beilin, said the Tulsa Police Department “deserves full credit for their impressive crime-fighting efforts — nowhere do we claim that the presence of Flock Safety technology is the sole factor in these public safety results.” 

Sign up to be notified whenever we publish a new story

The company sent a media release in January with the subject line: “TPD Clears 100% of Homicides Thanks to License Plate Readers.”

Johnston said the technology helps in other ways that aren’t represented in the data, such as removing violent criminals from the streets faster and helping officers confirm they have the right suspect.

Tulsa police mostly use the cameras to apprehend suspects and only occasionally use them to investigate or solve crimes, according to Johnston and court documents. In one case, Tulsa officers identified a vehicle leaving a murder scene with Flock cameras, using the data to help establish probable cause and arrest a suspect. However, the suspect was later found to be the wrong person. In other cases in Oklahoma, though, the cameras have proven undeniably useful — like when they helped capture a Chinese mobster who got all the way to Florida after he killed four people at a rural marijuana farm in 2022.

For violent crimes overall, Tulsa’s clearance rate hasn’t improved since the system was deployed in mid-2022, according to FBI data. Homicide, robbery, and property crime rates in the city and across Oklahoma have declined steadily over the past several decades, mirroring national trends.

Peer-reviewed studies show that license plate reader cameras have had little impact on solving or deterring crime in other U.S. cities. Jonathan Hofer, a researcher at the Independent Institute, a libertarian think tank, who has studied the effectiveness of license plate reader cameras, argues that Tulsa’s investment in the technology could have been better spent on more officers.

“With those millions, think about how many detectives you could have hired,” Hofer said. “There’s a possibility license plate readers increase certain types of crime because you’ve divested money that could have gone to police personnel.” 

Studies show that even hiring a single police officer can measurably reduce homicide and violent crime rates — especially in short-staffed departments like Tulsa, where budget constraints have kept salaries lower than those in competing agencies.

Johnston disagrees that the Flock system was a poor investment. He calls it a “force multiplier” that has made all 800 officers significantly more efficient. 

“Ten or so more officers? We’d love to have that … but you never know what benefit they’re going to have,” he said. “Having a resource available that supports the whole agency in the way that license plate reader data does, whether you’re an investigator or you’re a patrol officer responding to a call and finding a vehicle that’s leaving, it makes all of us much more efficient at our jobs.”

Tulsa pays Flock Safety nearly $700,000 annually through a no-bid contract, approved every year by the mayor. In 2023, the city spent over $2.5 million to build a real-time information center to monitor the data that runs software designed by Flock. The city also employs about 20 analysts and supervisors that staff the center, Johnston said. 

A longtime Oklahoma law states license plate reader data “shall not be used by any individual or agency for purposes other than enforcement of the ‘Compulsory Insurance Law.’” The Oklahoma Department of Transportation recently asked local agencies to remove Flock cameras from state-managed roads because of these legal concerns. Johnston said Tulsa avoided the issue by consulting with the Department of Transportation, which denied permission to install cameras on state property.

“When governments rely on unauthorized technology to surveil their citizens, policing methods become lazy,” said Gann. “It can’t replace the human element because you’re just capturing a moment in time. There’s no context to it, there’s no judgment to it.”

But Johnston said he believes that when license plate readers are combined with live-streaming cameras, or what he calls a “layered approach” to surveillance technology, they can reduce crime. Research on the effectiveness of real-time information centers and live-streaming cameras has yielded mixed results. 

Nationally, Flock Safety has aggressively marketed its technology and claims it has “helped solve 10% of reported crimes in the United States.” However, Flock’s internal research is not peer-reviewed and has been criticized. A co-author of Flock’s study says he “would have done things much differently.” 

Tulsa officials frequently appear in Flock’s marketing campaigns, press releases, and case studies. 

Johnston said Tulsa is featured in so many advertisements because it was early to adopt the company’s platform. He said the department does not receive discounts or incentives for participating in the advertising campaigns. He noted that one former Tulsa officer now works as a sales manager for Flock Safety but said there’s no major revolving door between the department and the company.

Flock Safety and its CEO say they aspire to “a future with a Flock camera on every street corner” and “an America where crime no longer exists.”

Privacy and legal concerns

Nguyen said because Flock cameras can continuously track cars, they may violate the Fourth Amendment of the U.S. Constitution, which protects against unreasonable search and seizure. Courts have ruled that attaching GPS trackers to cars constitutes a search requiring a warrant.

A Flock spokesperson disagreed.

“LPR cameras take photos of cars in public and cannot continuously track the movements of any individual,” said the spokesperson, Connor Metz. “Courts have persistently found that there is no reasonable expectation of privacy in a license plate on a vehicle on a public road, and photographing one is not a Fourth Amendment search.”

But Hofer said when governments take enough pictures of cars and are able, in practice, to use them to track individuals, this may violate the Fourth Amendment.

“If you use them in congruency with each other, you can determine travel patterns, speed of the car, obviously, direction. That seemingly poses a constitutional problem,” he said. “The line is up for the courts to decide … but I think it’s highly plausible that they get curtailed a little bit.”

Nguyen said the data is also often shared with other agencies, including the federal government. A federal grant from the U.S. Department of Justice paid for some of Tulsa’s first flock cameras in 2022, documents show. The Tulsa Police Department denied an open records request for a list of agencies it shares its data with. 

Johnston said he wasn’t sure how much access federal agencies have to the data but said it is shared with an intelligence fusion center in Oklahoma City, which aggregates local intelligence for federal use. A 2013 media report alleged the center shared information with an oil company and law enforcement that was used to surveil and disrupt environmental protesters opposing a pipeline project.

“We have significant concerns about who they share this data with,” Nguyen said. “Law enforcement agencies can say it stays within their agency, but Flock’s customer agreement that these agencies sign allows for the explicit sharing of the data captured to anywhere in the world.”

One of Tulsa’s contracts states: “Flock may access, use, preserve and/or disclose the Footage to law enforcement authorities, government officials, and/or third parties” if the company is “legally required to” or has a “good faith belief” it would be in the best interest to do so. 

The Tulsa Police Department’s Flock transparency portal states its data is stored for 30 days and “only shared with law enforcement agencies with reciprocal sharing.” Tulsa’s contract says it can search for cars in Flock’s “nationwide network.” The department says it prohibits use of the data for “immigration enforcement, traffic enforcement, harassment or intimidation, usage based solely on a protected class (i.e. race, sex, religion), or personal use.”

Public records show that Tulsa police receive automatic alerts flagging vehicles on national watchlists thousands of times each month. Nguyen said this raises concerns about unjustified stops, wrongful arrests, and the expanding reach of federal law enforcement. 

She pointed to the case of a U.S. veteran with no criminal record who was pulled over by Oklahoma City police five times in two months after Flock cameras repeatedly flagged his vehicle due to his inclusion in a federal terrorism database — a designation he had been contesting in court for years. In another case, an officer in Kansas was criminally charged for using the Flock system to stalk a former romantic partner. 

Gann said he’s heard from former Oklahoma police officers who retained access to their agency’s Flock platform on their phones even after leaving the department. Nguyen said she is aware of a case of an officer in Oklahoma who allegedly used a license plate reader system to track his ex-partner in a domestic violence situation.

A Flock spokesperson said “Flock’s platform requires a search justification for every search, providing a robust tracking and auditing capability for police leadership, city councils, and citizens.” Johnston acknowledged that misuse is possible but argued it’s no different than the potential abuse of any other police database.

Advocates also worry about the placement of these cameras in areas designated as “high-crime,” which are often minority communities. The Frontier submitted an open records request for the locations of all cameras currently operated by Tulsa police, but the department denied the request without explanation.

Flock cameras are used by about 60 law enforcement agencies in Oklahoma, according to the company. Oklahoma City had installed 90 cameras as of 2023.

Sen. Darrell Weaver, R-Moore, filed legislation this year to authorize the use of license plate readers in Oklahoma, but it didn’t get a hearing before a legislative deadline. A similar bill failed in the Senate last year by a 13-28 vote, despite Flock’s persistent efforts to sway lawmakers. The company has hired an Oklahoma lobbying firm and has urged law enforcement agencies to reach out to their representatives.

“We recommend that you call or email your elected representative in the Oklahoma Senate this week to express your support for the bill,” a Flock salesperson wrote in an email to agencies last year. “It’s critical that your legislator knows your Agency supports this measure and the importance of LPRs in helping law enforcement solve crime in your jurisdiction.”

Johnston said he hopes Oklahoma will pass new laws to keep up with the technology. 

“When that law was written, the focus was on the turnpikes and how license plate readers will be used on the turnpikes in an enforcement way,” Johnston said. “Now, we’ve seen changes with the technology being used differently, and it’s really up to our elected officials to take a look at what changes need to be made to help their communities best respond to crime.”