Police are stopping drivers on AI travel-pattern flags, and a Texas jury has sided with one who sued

Gizmodo reported this week that US law enforcement agencies are using automated licence plate readers and AI tools to pull over drivers who are not suspected of a specific crime. The agencies flag them for "suspicious" travel patterns instead. The warning came from the American Civil Liberties Union, which published a blog post on Monday built around Grace Knowles, a woman in her 20s from northern Virginia. On her way to Florida an officer stopped her for spending too much time in the left lane, questioned her about her travels and asked for proof she had been in New Orleans. He then identified himself as an FBI special agent working on a human and drug trafficking project and told her that her plate had been flagged for suspicious travel patterns. He showed her his screen. "It was basically a heat map of the U.S., and I could see different circles where I'd been, and when he clicked on it, it was a photo of my license plate," Knowles told the ACLU.

The flag came first, the suspicion came after

A plate reader used to answer a question an investigator already had. In the cases Gizmodo describes, the system asks the question. Flock Safety and Axon sell plate reader systems that scan vehicles as they pass and log the time and location. Now software reads the record for patterns and hands an officer a reason to stop someone.

The Knowles case does not stand alone. In March, The Markup reported that US Customs and Border Protection officers questioned a California grandmother about how often she visited local casinos. WIRED reported this month that Flock had built an AI investigative tool that "can identify drivers and track vehicles by their patterns of movement alone."

Look at where the judgement sits. The officer who pulled Knowles over had not reasoned his way to suspicion. A system had done that, and he acted on its output. In this setup the machine makes the inference and a human carries the consequences. AI does the routine work. You do the thinking. On this evidence, the thinking happened after the car had already stopped.

Pattern flags are cheap to switch on and expensive to defend

The pushback has reached juries and procurement desks. Earlier this month a federal jury in Texas sided with a man who sued over what his attorneys called an "unconstitutional traffic stop scheme" enabled by "AI-powered" plate readers. Police had pulled him over after a plate reader flagged his plate for driving from Houston to Carrizo Springs, staying overnight and returning home the next day. Agencies including the Los Angeles Police Department have suspended or ended their Flock contracts. An August YouGov survey found 46% of respondents opposed police using the technology in their communities, against 38% who supported it.

The same pattern runs well beyond policing. Fraud scoring, insider-risk tools and workforce analytics all flag people by pattern and pass the flag to someone with authority to act on it. If you are the person who acts, your judgement is the product. If you are the person who bought the system, its errors carry your name. Jay Stanley, a senior policy analyst at the ACLU, wrote that for many people these encounters "will be far more fraught, frightening and dangerous" than the one Knowles had.

Your Next Move

1. Map every pattern flag you touch. List each system in your work that scores people by behaviour, such as fraud, access, attendance or performance. For each one, write down who acts on a flag and what that person must check before acting.
2. Ask the vendor what triggered the flag. For any system you own or are about to buy, ask what pattern triggers a flag and how often it flags people who turn out to be innocent. A vendor that cannot answer is passing its liability to you.
3. Put your judgement before the action. If you act on machine output, record your own reasoning for each decision before you act. That record is the part of the job the system cannot do, and the part a jury will ask about.