5 Minutes
The knock on the door came while Angela Lipps was babysitting her grandchildren.
Minutes later, the 50-year-old Tennessee grandmother was in handcuffs, surrounded by U.S. marshals with guns drawn. The accusation sounded surreal: bank fraud in Fargo, North Dakota — a place Lipps says she had never visited, not even once.
What followed was a nightmare that lasted nearly half a year. And at the center of it all sat a familiar modern culprit: an AI-powered facial recognition match that investigators trusted far more than they should have.
Lipps, a mother of three and grandmother of five, has spent her entire life in north-central Tennessee, roughly 1,000 miles from Fargo. Yet in July 2025, authorities arrested her as a fugitive tied to a series of fraudulent bank withdrawals.
The reason? A surveillance video from a Fargo bank showing a woman using a fake U.S. Army military ID to withdraw tens of thousands of dollars. Detectives fed the footage into facial recognition software. The system flagged Lipps as a possible match.
That single lead appears to have carried enormous weight.
According to police files reviewed by local station WDAY, a detective compared the suspect in the video to Lipps and concluded the facial features, body type, and hair looked similar. No record suggests that investigators contacted Lipps beforehand to question her or confirm her whereabouts.
Instead, they moved forward with an arrest warrant.
When an Algorithm Becomes the Evidence
Because Lipps was treated as a fugitive from another state, she was jailed in Tennessee without bail. Days turned into weeks. Weeks turned into months.
For nearly four months, she remained locked in a county jail while the extradition process crawled forward. A court-appointed attorney eventually informed her that she would have to travel to North Dakota to contest the charges in person.
“I’ve never been to North Dakota. I don’t know anyone from North Dakota,” Lipps told reporters later.
The Fargo Police Department didn’t transport her immediately. In fact, it took 108 days after her arrest before authorities moved her from Tennessee to North Dakota. Only in December — more than five months after her arrest — did investigators conduct their first direct interview with her.
By that point, her defense attorney had already uncovered something striking: bank records placing Lipps more than 1,200 miles away in Tennessee at the exact time the fraud was happening.
In other words, the alibi investigators never bothered to check.
“If the only thing you have is facial recognition, you might want to dig a little deeper,” said Jay Greenwood, the lawyer representing Lipps.
Once the records were presented, the case collapsed almost immediately. Authorities dropped the charges and released Lipps from jail on Christmas Eve.
Freedom, however, came with a bitter twist. She had no money and no way to get home. According to WDAY, local defense attorneys pooled funds for a hotel room while a nonprofit organization helped arrange transportation back to Tennessee.
Lipps says the ordeal cost her nearly everything — her home, her car, even her dog.
“I had my summer clothes on, no coat,” she recalled of her release in North Dakota’s winter cold. “There was snow on the ground. I was scared. I just wanted to go home.”
She also says no one from the Fargo Police Department has apologized.
Her case is not an isolated one. In recent years, multiple U.S. police departments have faced scrutiny after facial recognition systems produced false matches that led to wrongful arrests. In 2024, New York police arrested a man based on grainy CCTV footage despite him being significantly taller than the suspect. In Detroit, a woman filed a lawsuit after authorities arrested her for murder following an AI identification that overlooked obvious physical differences.
Facial recognition technology is spreading quickly across law enforcement agencies, often promoted as a powerful investigative tool. Yet experts continue to warn about false positives, especially when systems analyze low-quality images or when investigators treat algorithmic suggestions as confirmation rather than clues.
For Angela Lipps, the consequences were devastatingly real — a reminder that when artificial intelligence enters the justice system, a single digital mistake can still ruin a human life.
Source: inforum
Leave a Comment