Every Car, Everywhere, All at Once—Part 2: License Plate Readers Are on a Collision Course With the Constitution
This is the second installment in a three-part series on automated license plate reader technology.
Read the introduction and see other posts here.
Last month, a federal judge in Oklahoma became first in the country to rule that a Flock automated license plate reader (ALPR) search violated a woman’s Fourth Amendment rights.
Flock’s network of over 120,000 cameras across the United States picked up Melissa Kyle’s SUV more than 50 times over the course of a month, allowing Tulsa County Sheriff’s Deputy Freddie Alaniz to trace her from Missouri through Oklahoma and toward the California-Arizona border and back. U.S. District Judge Sara E. Hill suppressed the search (which turned up 91 pounds of methamphetamine), reasoning that compiling retroactive location records is “a type of indiscriminate mass surveillance” categorically different from observing a single plate at a single moment. “[T]he Fourth Amendment requires courts to draw a line when the cost is too great,” she wrote.
While the ability to reconstruct the route of any driver on American roads certainly introduces novel risks, the core constitutional question is an old one: What is a “search” under the Fourth Amendment, and when do police need to obtain a warrant? The U.S. Supreme Court has already answered that question for GPS trackers and cellphone records, and the Oklahoma ruling is the first sign that those precedents apply to ALPRs too. Part 1 of this series explored the public safety case for ALPRs; Part 2 will take stock of what can go wrong and examine case law as it catches up with the technology.
What Are the Risks of ALPRs?
Just before 10:00 PM on Oct. 4, 2025, a Dodge Durango traveling over 100 miles per hour struck a motorcycle and a Ford Focus on Interstate 4 in Florida, killing three people. The driver fled. Searching Flock’s network, investigators found a black Durango nearby registered to 23-year-old Lindsey Isaacs. Charged with eight felonies—including three counts of vehicular homicide—Isaacs spent 13 days in jail before a second team of investigators noted a crucial detail: The SUV she had been driving that day did not have a scratch on it.
The failure in this case was not technological; rather, it was profoundly human. Once the system flagged a plausible vehicle, investigators trusted the computer at the expense of their own common sense. Ultimately, the charges against Isaacs were dropped and another woman was charged with the crime. But this case exposes an overlooked risk of artificial intelligence (AI) use within the criminal justice system: automation bias. The more advanced the system, the more likely people are to believe the machine over their own eyes.
One measure of the system’s power is the sheer number of officers willing to risk their careers to abuse it. The Washington Post documented at least 50 officers who have been accused, charged, or convicted of abusing the system, mostly to stalk ex-partners. In Georgia, one of the few states with an ALPR-specific misconduct law, at least 20 people— including a police chief charged with stalking, harassment, and ALPR misuse—have been arrested. And in Indiana, where several ALPR bills have failed to become law, an officer was found to have performed 3,759 searches on vehicles used by his wife and two others over a 10-month period. In the absence of meaningful oversight, individual officers have treated internal controls as a formality at best and with ridicule at worst. Public records requests show that officers routinely enter things like “LMAO” or “idk” into the fields intended to document the rationale for a search, apparently quite certain no one is checking the logs.
Nationwide data sharing between states with different laws has led to additional legal grey areas. In 2025, a Texas sheriff’s office searching for a woman who allegedly self-administered an abortion queried more than 80,000 ALPRs around the country—including cameras in Illinois, where state law restricts the use of such data for abortion investigations. Once networked, ALPR data collected in one location for one purpose tends to leak across jurisdictional and regulatory bounds. For example, Florida wildlife officers have conducted searches on behalf of federal immigration authorities, and Flock has acknowledged giving pilot access to four federal agencies. Even when the government does nothing with the data, simply knowing that it can track visits to protests, churches, clinics, gun shops, and addiction meetings could have a chilling effect on constitutional freedoms.
To its credit, Flock has responded to the backlash with real changes. Over the summer, the company cut its default data retention from 30 days down to seven, although it can be preserved for longer when tied to a case number. It also required agencies to activate an internal audit tool that flags abnormal search patterns, which police in Savannah, Georgia, credit with identifying officer misuse. Agencies also gained finer control over outside access, including the ability to restrict outside department searches by investigation type.
But the technology carries a more subtle risk that no company policy can mitigate—namely, the temptation to use it for everything. In the past, an officer had to notice an expired tag. Now, plate readers can identify many expired tags at the same time. For example, the city of New Haven, Connecticut, pays a private contractor to patrol its streets with plate readers, hunting for cars whose owners owe back taxes. The firm, which keeps a 38 percent commission on every vehicle it adds to the tax rolls and charges owners a $50 fee on top, recovered nearly $24 million for the city between 2024 and 2025. According to procedural justice theory, cooperation with police depends on whether people see them as fair, and a network that turns every expired tag into an automatic fine is likely to be seen as a revenue generator rather than a public safety tool.
Are ALPRs Constitutional?
ALPRs went undefeated in court for years. Flock and the government argue that license plates exist precisely so the state can identify vehicles on public roads, reasoning that if courts permit an officer to search any plate, for any reason, without a warrant, then a computer should be able to as well. Judges in Massachusetts, Virginia, Washington, and elsewhere have agreed. A federal judge upheld Norfolk, Virginia’s 170-camera network on those very grounds in January 2025, finding it does not capture “the whole of a person’s movements.” By Flock’s count, more than 30 state and federal decisions have reached the same conclusion.
The recent Oklahoma ruling broke that streak, citing a decade of Supreme Court precedent that requires police to obtain a warrant in order to track vehicles over an extended period. In United States v. Jones (2012), the court found that attaching a GPS tracker to a suspect’s car for 28 days qualifies as a search, even when the car never leaves public roads. Carpenter v. United States (2018) extended the principle to digital records, holding that a detailed, retrospective log of a person’s movements requires a warrant to obtain. In 2021, the Fourth Circuit applied Carpenter to aerial surveillance, ruling that Baltimore’s spy-plane program amounted to a warrantless search of residents’ movements.
In Chatrie v. United States (2026), the Supreme Court held 6-3 that police conducted a Fourth Amendment search when they obtained cellphone location data identifying devices near a Virginia bank robbery. Writing for the majority, Justice Elena Kagan explained that individuals retain a reasonable expectation of privacy in records of their location, even when police request only a limited time window and even when a private company holds the data. Fourth Amendment scholars believe Chatrie has shifted the constitutional test from whether police actually obtained the whole of a person’s movements to whether the system they queried is capable of doing so.
Conventional traffic cameras passed constitutional muster because they are narrow in scope, triggered by a specific violation at a single location, at a single moment in time. With ALPRs, the difference is the network. While a single journey is public, the compiled record of a person’s day-to-day movements is not. As ALPR networks become more capable of reconstructing a person’s movements, they more closely resemble forms of surveillance the Supreme Court says constitute a search under the Fourth Amendment. Currently on appeal before the Fourth Circuit, Schmidt v. City of Norfolk will be the next test to determine whether ALPRs cross that line.
Beyond the Courtroom
After her 13 days in jail for a crash she had nothing to do with, Lindsey Isaacs had one request for the Senate Judiciary Committee investigating ALPRs: to create minimal safeguards so that no one else goes through what she did. Most Americans would agree—although they want police to have new tools to recover stolen cars and find missing children, they do not want to be tracked everywhere they go. That balance will not come from the courts alone; it must be written into law.
The final installment of this series will explore the state statutes already on the books and the federal proposals now circulating in Congress, offering practical recommendations for legislators as they consider ALPR bills in 2027 and beyond.