Tennessee Grandmother Jailed for Months Due to AI Facial Recognition Mistake

Facial recognition technology has once again come under scrutiny after a Tennessee grandmother was wrongfully arrested and detained for nearly half a year. Angela Lipps, 50, was taken into custody by US Marshals at gunpoint last July while babysitting, following an incorrect match made by AI software.

A Case of Mistaken Identity
Authorities in Fargo, North Dakota, initiated the arrest based on an investigation into bank fraud that occurred in April and May 2025. Detectives utilized facial recognition software to analyze surveillance footage of a suspect using a fake US Army ID to withdraw large sums of money. The system flagged Lipps as a potential match after comparing the footage with her Tennessee driver's license and social media images.

Lipps was charged with theft and unauthorized use of personal identifying information. She spent 108 days in a Tennessee jail awaiting extradition, unable to post bail. Her attorney, Jay Greenwood, later obtained bank records that definitively proved Lipps was in Tennessee—over 1,200 miles from the crime scene—purchasing cigarettes and depositing government checks at the time of the alleged offense.

Consequences and Aftermath
Following the verification of her alibi, Lipps was released on Christmas Eve. The ordeal left her stranded in Fargo without resources. Local attorneys and a non-profit organization, the F5 Project, provided her with shelter and assisted in arranging her return home.

The financial and personal toll of the wrongful imprisonment was severe. Lipps reported losing her home, her vehicle, and her dog because she was unable to cover her bills while incarcerated. According to reports, the Fargo Police Department has not issued an apology for the error.

Systemic Issues in AI Policing
This incident adds to a growing list of wrongful arrests caused by reliance on facial recognition technology without sufficient human verification. The case draws parallels to that of Robert Williams, a Detroit man wrongfully arrested in 2020 based on a similar algorithmic match. Following Williams' lawsuit, Detroit amended its policies to prevent arrests based solely on facial recognition results.

Experts warn that these technologies often exhibit higher error rates when analyzing specific demographics and emphasize that AI findings should serve only as investigative leads, not definitive evidence. As AI integration in policing expands, advocates call for stricter oversight to prevent future miscarriages of justice.