What began as an ordinary summer day for a Tennessee grandmother quickly turned into a life-altering nightmare after artificial intelligence technology wrongly identified her as a suspect in a crime hundreds of miles away.

Angela Lipps, 50, says she spent nearly six months in jail after facial recognition software incorrectly connected her to a bank fraud investigation in Fargo, North Dakota—a place she insists she had never visited. As reported by The Independent, the case eventually collapsed when evidence proved she was in Tennessee at the time of the alleged crimes.

But by then, Lipps says the damage had already been done.

The ordeal began on July 14 while Lipps was babysitting four children at her Tennessee home. According to her account, U.S. Marshals arrived and arrested her at gunpoint after authorities identified her as a suspect in an organized bank fraud scheme involving stolen identities.

Lipps was taken to a local jail as a fugitive from justice tied to charges filed in North Dakota. Because of that designation, she was held without bail while officials arranged for her extradition.

“I’ve never been to North Dakota,” Lipps later said, explaining that her travel history had been limited to nearby states and that she had never even flown on an airplane.

The investigation that led to her arrest began months earlier when Fargo detectives were looking into several fraudulent bank withdrawals. Surveillance footage showed a woman allegedly using a fake military identification card to withdraw tens of thousands of dollars from bank accounts.

Police ran the images through facial recognition software, which returned a match: Angela Lipps.

Investigators then compared the footage with Lipps’ Tennessee driver’s license photo and reviewed her social media images. Based on similarities in facial features and appearance, authorities moved forward with charges.

Lipps says she was never contacted by investigators before the arrest warrant was issued.

She remained jailed in Tennessee for more than three months while extradition proceedings played out. Eventually, she was transported more than 1,200 miles to Fargo to face charges that included identity theft and unauthorized use of personal information.

Her court-appointed attorney soon uncovered evidence that challenged the case. Bank and transaction records showed Lipps was actively making purchases in Tennessee at the same time the alleged crimes were happening in North Dakota.

Records indicated she had been depositing Social Security checks, buying food, and making everyday purchases locally during the time investigators believed she was committing fraud.

Once those records were reviewed, prosecutors dropped all charges just days before Christmas, and Lipps was released.

The dismissal, however, left her stranded in Fargo in winter weather with little money and only summer clothing. Local advocates helped her find a place to stay and eventually assisted her in traveling back home.

Lipps says the experience upended her life. While she was jailed, she was unable to pay her bills and ultimately lost her home, her car, and even her dog.

The bank fraud investigation itself remains open, but no suspects have been arrested.

Meanwhile, the case has become one of several documented wrongful arrests in the United States tied to facial recognition technology—raising new questions about how AI is used in criminal investigations and the potential consequences when it gets it wrong.