Before the U.S. marshals put Angela Lipps on a plane, she had never been on one. It’s difficult to get rid of that detail that keeps coming up in discussions about her case. One afternoon in July of last year, a fifty-year-old grandmother from Elizabethton, Tennessee, was watching four kids in her living room when she unexpectedly saw guns at her front door. After being incarcerated for almost six months, she was released by Christmas Eve. However, she was left stranded in North Dakota without a way home and was staying in a hotel that was funded by strangers while eating food that had been donated.
She is now suing, and her lawyer is telling everyone who will listen that this is a unique lawsuit. The police departments involved, the developers of the facial recognition software, and, according to documents circulating among civil rights lawyers, a larger effort to bring AI companies themselves into court are among the defendants. The legal theory that is emerging is the one Silicon Valley has been secretly fearing for years: that someone upstream must be held accountable when an algorithm causes actual, quantifiable harm, regardless of whether OpenAI is named directly or drawn in through related claims.
Speaking with those who follow these cases, it seems like Lipps might not be the best plaintiff for the tech sector to deal with. She is not a criminal, a hacker, or someone with a convoluted past that a defense team could dissect. She is a grandmother of five who, while incarcerated in Tennessee, lost her house, car, and dog because, according to the police department, a piece of software operated by a partner agency in West Fargo called Clearview AI identified her face as a match. That’s all. The lead was that.
Jay Greenwood, her attorney, made a statement to local reporters that has been cited by others in the legal community on several occasions. “If the only thing you have is facial recognition, I might want to dig a little deeper.” It sounds almost soft. It isn’t. If you read it carefully, it’s a critique of the entire investigative culture that has developed over the last five years, where probable cause increasingly refers to whatever the program generates.
At a press conference, the Fargo police chief acknowledged “a few errors.” He did not offer an apology. Attorneys representing plaintiffs in this case consistently emphasize that above all else. As of yet, the system does not incorporate regret. A protocol that views an AI match as a lead instead of a conclusion does not exist. And juries, increasingly, don’t like that.
It’s worth remembering Lipps isn’t the first. There was the Baltimore high school student last October whose bag of Doritos got read as a firearm by an AI surveillance system. The man in the UK arrested for a burglary in a city he’d never visited because face-scanning software couldn’t distinguish him from another person of South Asian heritage. The lawsuits continue to mount, and the pattern continues to recur. The courts’ willingness to hear them is what is evolving.
The case is already making progress, regardless of whether Lipps prevails completely or reaches a quiet settlement. Facial recognition partnerships are reportedly being reviewed by police departments across the nation. According to reports, insurance underwriters covering AI vendors are posing more pointed queries. And in the kind of irony that journalists are trained to notice but never quite get used to, a woman who’d never flown on a plane until the government put her on one in handcuffs is now at the center of a fight that could reshape how the most powerful technology companies in the world handle being wrong.
It’s difficult not to support her. Predicting what will happen next is even more difficult.
