Angela Lips, a 50-year-old woman from Tennessee, never imagined that a misidentification by AI facial recognition would leave her locked up in jail for nearly half a year. This case is not only the ninth recently reported instance of wrongful identification by facial recognition technology in the United States, but it also exposes the multiple loopholes behind the misuse of this technology in law enforcement.
From April to May last year, police in Fargo, North Dakota, were investigating a bank fraud case: a woman had used a fake military ID to withdraw nearly ten thousand dollars. After inputting surveillance footage into the facial recognition system, the system matched Angela Lips’ information.
What’s astonishing is that Lips had never set foot in North Dakota and was over 1,200 miles away in Tennessee at the time of the crime.
Even so, a detective identified Lips as the perpetrator solely based on facial features, body shape, and hairstyle. On July 14, while Lips was looking after four children, U.S. Marshals—without ever having contacted her before—arrested her at gunpoint. For the next 108 days, Lips was held in a county jail in Tennessee until North Dakota police took over the case.
Her lawyer, Jay Greenwood, immediately requested bank records, which became the key evidence to overturn the case. On December 19, five months after Lips was arrested, Fargo police finally met with her and her lawyer.
The bank records clearly showed that during the time of the crime, Lips had been in Tennessee purchasing cigarettes and regularly depositing her social security checks, making it impossible for her to have committed the crime in North Dakota.
On Christmas Eve, the case was officially dismissed. However, Lips was left penniless, without a coat, unable to return home, and ended up losing her house, her car, and even her pet dog.
This is not an isolated incident. A January 2025 investigation by The Washington Post revealed that there have been at least eight similar wrongful arrests due to facial recognition misidentification in the U.S. In every case, investigators skipped basic steps such as verifying alibis or cross-checking physical characteristics, and acted solely on the technology's match result.
Now, Lips has returned to her home in Tennessee, but is still waiting for an official apology from the Fargo Police Department.
This case once again brings to light the issue of abuse of AI facial recognition technology in law enforcement. The gap between technological accuracy and the rigor required in law enforcement urgently needs to be bridged.