Boren leads Senate discussion on Flock cameras, government surveillance

Oklahoma Senators gathered Wednesday for an interim study on Flock cameras and automated license plate readers requested by Sen. Mary Boren, D-Norman.

Boren called for the interim study to examine privacy concerns surrounding the growing use of automated license plate readers and the collection and storage of people's data. On Oct. 1, in United States v. Kyle, a case involving a Tulsa County deputy searching two license plate reader systems and retrieving more than 50 records of a woman's movements across multiple states, an Oklahoma federal judge ruled that searching a driver's monthlong travel history without a warrant violated the Fourth Amendment.

"Do Oklahomans have a right to know where government surveillance cameras are placed in our communities?I think that's an issue that we're facing," Boren said in her introduction. "From vendor contracts, corporate profits to agency expanding authorities, there are complex incentives and power plays, and we cannot abandon a deliberate principled process. And, Mr. Chair, that's what I'm hoping that this interim study will do."

Remarks on Flock cameras and automated license plate readers were provided by Rep. Tom Gann, R-Inola; Northeastern State University criminal justice professor Michael Wilds; and Chad Marlow, senior policy counsel for privacy and technology at the American Civil Liberties Union.

Michael Wilds helped provide the legal background, beginning with the reasonable expectation of privacy established in the 1967 case Katz v. United States. He also described the 2012 case United States v. Jones, regarding GPS devices attached to vehicles. The court rejected this use of GPS tracking because the car was private property and raised concerns about tracking individuals over time. He also described cases where courts deemed warrantless searches of cellphone contents and the use of thermal imaging technology to examine homes unconstitutional.

He explained how these precedents relate to the court's decision in United States v. Kyle, but said the ruling raised new questions about how those protections apply to license plate reader data and what restrictions lawmakers should consider, such as how long records are retained, whether records from different camera systems can be combined, who can search the information and for what reasons, and who should audit access and prevent misuse.

He also said Oklahoma has the ability to provide greater constitutional privacy protections than federal law, which Rep. Gann expanded on in his presentation.

"I'm going to compare the statutory question to the constitutional question that was brought out by Judge Hill's ruling, and that distinction leads to today's central question. When does observing a plate become a travel history?" Gann said.

Gann said uncertainty arose with the passage of Title 47, Section 7-606.1, of Oklahoma law in 2016, which allows the use of license plate reader data for compulsory insurance enforcement, "or as otherwise permitted by law." Gann played a recording from the legislative consideration of the measure, in which Trent Baggett, then representing the District Attorneys Council, said, "All that we're looking to use that for is uninsured motorists. It's not to look for ... anything beyond that, that's not what this is designed for."

Gann said since hosting interim studies on similar topics in 2024 and 2025, he had been contacted by multiple people "who had been ... detained by law enforcement people at gunpoint, because of a misread in a license plate."

"The last person that talked to me in '25, who was actually an adjunct professor from OU, that had this happen to him ... he said it just really has changed the way he even drives. When he sees the police car in the rearview mirror, it just changes him. It changes what he does, how he thinks, and that's sad," Gann said.

He distinguished observing a license plate on a public road from maintaining a searchable history of everywhere a vehicle has traveled, and suggested altering policy around tracking to plainly prohibit government use of ALPRs for persistent tracking and surveillance.

"A license plate may be visible in the public, but a monthlong, multi-state searchable database of a person's vehicle movement is not plain view," Gann said.

Marlow, of the ACLU, provided more restrictive proposals for addressing government tracking and surveillance. He leads the ACLU's national "Get the Flock Out" campaign, and unlike Gann's proposal to prohibit persistent tracking and surveillance, the ACLU wants government use of ALPRs banned entirely.

"We understand fully that while guardrails can make ALPR use somewhat safer, they cannot make them safe," said Marlow.

He explained that ALPRs don't just take a picture of a license plate; each time an ALPR photographs a vehicle or reads a license plate, it also records the date, time and location of each observation and vehicle characteristics such as make, model and color, down to scratches and bumper stickers. Those records are then checked against law enforcement databases, and modern systems connect records from cameras across the country rather than limiting access to one department's local collection.

The implication of this, he said, is that the government gains the ability to track people's locations and movements over time.

"What places they visit throughout their day, and, with the increasing help of AI, can analyze past behavior and even attempt to predict future behavior. This can provide the government with access to some very personal and intimate details of your life such as what religious institutions you attend, what political meetings you go to, what doctors' offices, like a psychiatrist or oncologist you visit … The list goes on and on," Marlow said.

He added that Flock alone collects approximately 20 billion license plate scans per month, which is too much data for humans to analyze without AI.

"Not only can AI easily review billions of plate scans, but their natural language models allow AI to be instructed to detect everyone's patterns of life, and can even use ALPR data to manufacture criminal suspicion," Marlow said.

He cited a Florida case where a woman was jailed for 13 days, three of those in solitary confinement, for a multiple vehicular homicide crash Flock misidentified her SUV as being involved in.

He also noted a recent study of FBI crime data in Atlanta that looked at case closure rates before and after 5,000 Flock ALPRs were implemented across the metro area and found no change. Instead, he said it found that those cars, or that wanted person, would have been found anyway. Marlow also questioned the accuracy of the technology.

"Earlier this year, the LAPD canceled its Flock contract after its own Inspector General learned that one in three cars that were identified by a Flock ALPR in Los Angeles as being stolen were incorrectly identified," he said.

If ALPR searches are permitted with guardrails, Marlow suggested narrowing the searches to vehicles registered to individuals with outstanding felony warrants, vehicles involved in or connected to felonies and vehicles associated with missing persons and stolen vehicles. He said that ALPR data sharing should only occur between state and local agencies, and that the ACLU recommends a 48-hour retention period for data that is not connected to a specific criminal investigation.

"Without the legislature's intervention, things are only going to get worse," Marlow said.

Boren thanked Public Safety Committee Chair Warren Hamilton, R-McCurtain, for agreeing to hear the study.

"We must balance Fourth Amendment privacy protections with legitimate law enforcement needs," she said. "Unless the law is clarified, defense counsel may identify a significant weakness in the use of that data. This discussion needs to continue so that we clarify the permitted use of ALPRs under Oklahoma law as well as constitutional law."