A Tennessee grandmother, Angela Lipps, spent nearly six months in jail after a facial recognition error tied her to a Fargo, North Dakota bank fraud case, even though she had never been to North Dakota.

Lipps was arrested at her Tennessee home while babysitting, then extradited to North Dakota and charged with unauthorized use of personal identifying information and theft based on AI-generated facial recognition matches.

Bank records later showed Lipps was at her Tennessee home during the alleged fraudulent activities, contradicting the Fargo police’s accusation and prompting officials to interview her only after her lengthy detention.

The case spotlights broader concerns about AI facial recognition errors and misidentifications by law enforcement, fueling calls for more robust verification beyond facial matches.

Lipps’s lawyer argued that facial recognition should not be the sole basis for prosecution, and a nurse-authored timeline suggested a significant misidentification.

Lipps was released after her attorney obtained bank records establishing her alibi; she returned home but lost her home, car, and dog due to unpaid bills accrued during detention and received little formal apology.

The Fargo police investigation into the bank fraud continued with no arrests at the time of Lipps’s release, and authorities faced scrutiny for not interviewing her earlier.

Investigators used facial recognition on surveillance footage of a woman using a fake military ID to withdraw money and concluded Lipps resembled the suspect, leading to her identification.

Lipps was released on Christmas Eve after the case was dismissed, but she remained stranded in Fargo until local defenders and the F5 Project helped her return home, having lost her home, car, and dog due to detention-related expenses.