Police across the United States are increasingly pulling drivers over not because of anything they did on the road, but because an artificial intelligence system decided their travel patterns looked "suspicious," according to a new warning from the American Civil Liberties Union.
In a post published this week by Jay Stanley of the ACLU's Speech, Privacy and Technology Project, the civil-liberties group argues that AI-powered license plate surveillance has crossed a line: from helping investigate specific crimes to manufacturing suspicion where none previously existed. For more context on this story, see our ongoing latest AI developments.
"Anybody is susceptible to being stopped, searched, or worse," the post's subtitle warns. "AI makes new forms of abuse possible."
A Traffic Stop That Wasn't About Traffic
The ACLU's post centers on the experience of Grace Knowles, a woman in her twenties from northern Virginia who is sharing her story publicly for the first time.
In September 2025, between jobs, Knowles set out on a road trip that took her through Oregon, Texas and Louisiana. The day after she left New Orleans, bound for Florida, a police officer pulled her over despite, she says, her driving at a normal speed.
The stated reason was flimsy: he claimed she had been "in the left lane too long." Then the questions turned personal — where she was coming from, where she was going. When she said she had driven from New Orleans, he demanded proof. She offered a pamphlet from a whiskey museum she had visited.
Only then did the officer reveal the real reason for the stop. "I just had to be sure, because I'm not a regular cop, I'm part of the FBI, and when I scanned your plate I could see you had suspicious travel patterns," he told her, she recounted to the ACLU.
He even suggested she was lying about New Orleans because his system "didn't show any hits there." To reassure her, he invited her into his patrol car and showed her the software: a heat map of the United States dotted with circles marking everywhere she had been. Clicking a circle pulled up a photograph of her license plate.
The officer described himself as a special agent and said plates are scanned "everywhere" — including, he claimed, mounted on the back of highway road signs where drivers rarely notice them.
The actual source of the stop, per the ACLU's account: a surveillance system with access to automated license plate reader (ALPR) data had flagged her cross-country driving pattern as suspicious. No crime, no warrant, no individualized suspicion — just an algorithm's judgment about where a young woman had driven.
From Recording Plates to Generating Suspicion
License plate readers themselves are not new. What has changed is what is done with the data they collect.
In July 2025, the ACLU first flagged that Flock — the dominant player in ALPR surveillance — was touting an AI capability that went beyond matching plates against hotlists. The company boasted its system could alert police to drivers whose travel patterns were "suspicious," and could even map people's associations by analyzing which vehicles were frequently seen together.
Stanley called that a significant leap: using ALPR data to investigate a suspicious vehicle is one thing; generating suspicion out of ordinary driving is another.
A year later, Wired obtained detailed files describing Flock's next-generation AI system, already in use by some police departments. Those files revealed the capability for wide-ranging, open-ended AI queries against license plate data — and not only plate readings. The system could also draw on arrest records, case files and databases from the commercial data broker industry, which the ACLU describes as packed with personal information on nearly all Americans.
In other words, an officer can now ask open-ended questions of an AI that fuses where you have driven with a sweeping trove of third-party data, and get back a lead — no tip, crime report or judicial process required.
'Whisper Stops' and 'Intel Stops'
The pattern Knowles experienced has a name in law enforcement circles. The ACLU says stories from across the country are emerging of police conducting "pretextual stops" — pulling drivers over for a claimed, often invented, minor infraction such as drifting out of a lane or failing to signal, as cover for the real reason for the stop.
Police themselves have names for these detentions: "whisper stops" or "intel stops."
Pretextual stops are not inherently new — traffic code is vast, and officers have long had discretion over which violations to enforce. What AI changes is the targeting. Where an officer once relied on a hunch, a hotlist hit or a drug-interdiction heuristic, an algorithm can now mine months of a person's movements and hand back a list of candidates for a stop.
Civil-liberties groups argue that inverts the purpose of surveillance infrastructure. Systems sold to the public as tools to recover stolen cars or find missing people are being used to generate suspicion about people who have done nothing but drive.
Why the ACLU Is Sounding the Alarm Now
Stanley's post frames the trend as a systemic risk rather than a collection of isolated incidents. Because ALPR networks blanket highways and city streets, and because the AI analysis layers are new and lightly regulated, the population of people exposed to these stops is effectively everyone with a license plate.
"The growth of AI license plate surveillance from Flock and other companies means that more and more innocent people may be stopped in this way," Stanley wrote — noting that most of them will not get the impromptu software demonstration Knowles received.
The Knowles story illustrates the asymmetry of the encounter. She was questioned, asked to produce proof of her own whereabouts, and only learned the reason for the stop because the officer volunteered it. A traveler with less patience, or a darker room in the data, could fare far worse.
What Happens Next
Flock's AI-driven analysis is already in operational use, according to Wired's reporting on the company's internal files, and the ACLU's post suggests usage is spreading rather than receding. The group is calling public attention to the practice as courts, state legislatures and city councils continue to weigh how — and whether — automated license plate data should be governed.
For now, the advice implicit in the ACLU's warning is stark: if your driving pattern looks unusual to a machine, a minor traffic excuse may be all it takes for the stop to begin.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
