A Grandmother Lost Five Months to a Facial Recognition Error

Security surveillance camera mounted on a concrete wall in harsh lighting — Photo via Unsplash

TL;DR: Angela Lipps, a Tennessee grandmother, was arrested at gunpoint by U.S. Marshals in July 2025 for bank fraud committed in Fargo, North Dakota, a state she’d never visited. West Fargo Police used Clearview AI to flag her as a “potential suspect.” A Fargo detective looked at her social media, decided she matched, and filed charges. Lipps spent five months in jail. Her bank records proved she was home in Tennessee the entire time. The charges were dismissed on Christmas Eve 2025. She lost her home, her car, and her dog. Police haven’t apologized. She’s at least the 13th person wrongfully arrested because of facial recognition.

Arrested at Gunpoint While Babysitting

In July 2025, U.S. Marshals showed up at Angela Lipps’ Tennessee home with guns drawn. She was babysitting four young children at the time ^[1]. They told her she was wanted for bank fraud in North Dakota: four counts of unauthorized use of personal identifying information and four counts of theft ^[2].

Someone had used a fake U.S. Army military ID to withdraw tens of thousands of dollars from banks in the Fargo area between April and May 2025 ^[2]. The Fargo Police Department was investigating. They needed a suspect.

Here’s how they got one: Fargo PD doesn’t have its own facial recognition system. But neighboring West Fargo does, and they use Clearview AI, the startup with a database of billions of photos scraped from the internet, including social media ^[1]. West Fargo ran the suspect’s image through Clearview. It returned Angela Lipps as a “potential suspect.”

A Fargo detective then looked at Lipps’ social media accounts and Tennessee driver’s license photo. He wrote in his charging document that she “does appear to be the suspect based on facial features, body type and hairstyle and color” ^[1]. That’s it. A computer said maybe. A cop said yeah, looks right. A grandmother lost everything.

Five Months. No Bail. No Investigation.

Because Lipps was classified as an alleged fugitive (arrested in Tennessee on a North Dakota warrant) she was held without bail ^[1]. She sat in a Tennessee jail for nearly four months before being transferred to North Dakota.

During those months, nobody checked if she’d actually been in Fargo. Nobody pulled her bank records. Nobody looked at her phone location data. The simplest alibi check in the world (was she actually there?) never happened before they locked her up.

An internal email later revealed that six Fargo detectives were notified of Lipps’ arrest in July 2025, nearly five months before police claimed to know about her custody ^[3]. They knew she was sitting in jail. They sat on it.

Meanwhile, Lipps couldn’t pay her bills. She lost her home. She lost her car. She lost her dog ^[1]. Five months of her life, destroyed because an algorithm said she looked like someone.

Bank Records Proved She Was 1,200 Miles Away

Her attorney, Eric Rice, did what the Fargo Police Department apparently couldn’t be bothered to do: he pulled Lipps’ bank records ^[3]. They showed she was home in Tennessee, 1,200 miles from Fargo, at the exact times the fraud was being committed.

The charges were dismissed on Christmas Eve 2025 ^[3]. Five months after her arrest. After she’d already lost her housing, her transportation, and her pet.

But here’s the kicker: the charges were dismissed “without prejudice.” That means Lipps technically remains under investigation despite exonerating evidence ^[3]. Fargo Police have not apologized and claim their investigation is “ongoing” ^[3].

“I haven’t seen any evidence in the file that Angela is connected to this in any way,” Rice told reporters ^[3]. “I hope Angela’s matter can serve as a warning to law enforcement to not use AI as a shortcut.”

The 13th Known Victim. Probably Not the Last.

Lipps is at least the 13th person in the United States wrongfully arrested because of facial recognition technology, according to the ACLU ^[4]. The actual number is almost certainly higher. Many victims never learn that facial recognition triggered their arrest, and police departments aren’t required to disclose when they use it.

Here’s the running list of known cases ^[4]:

Robert Williams: Detroit, January 2020. First publicly reported wrongful arrest from facial recognition.
Michael Oliver: Detroit, 2020.
Nijeer Parks: New Jersey, 2019 (case became public later).
Randal Quran Reid: Georgia, November 2022. Arrested for crimes in Louisiana, a state he’d never been to. Held nearly a week.
Porcha Woodruff: Detroit, February 2023. Eight months pregnant when arrested.
Jason Killinger: Reno, Nevada, September 2023.
Robert Dillion: Jacksonville Beach, Florida, August 2024.
Javier Lorenzano-Nunez: Phoenix, Arizona, October 2024.
Trevis Williams: New York City, April 2025. We covered his case.
Angela Lipps: Tennessee/North Dakota, July 2025. Five months in jail.
Beau Burgess: Orlando, Florida, August 2025.

Notice the acceleration. Three cases in 2020–2022. Three in 2023–2024. Five in 2025 alone. NBC News reports that “facial recognition policing errors are on the rise” ^[5], not because the technology is getting worse, but because more departments are deploying it with less oversight.

The Clearview AI Problem

Clearview AI shows up in case after case. Its database contains billions of photos scraped from Facebook, Instagram, LinkedIn, and other platforms: photos people posted without any expectation they’d end up in a police lineup ^[1].

Clearview’s own contracts acknowledge that officers “must conduct further research in order to verify identities or other data generated by the system” ^[4]. Translation: we know this thing gets it wrong. Don’t just arrest whoever it spits out.

But that’s exactly what keeps happening. A detective gets a Clearview hit, glances at a social media photo, decides “yeah, looks right,” and files charges. No corroborating evidence. No alibi check. No second opinion. The algorithm becomes the investigation.

Multiple studies show facial recognition systems perform worst on darker-skinned faces, women, and older adults ^[4]. The technology’s error rates aren’t evenly distributed. They fall hardest on the people least able to fight back.

What You Can Do

Know your rights. If you’re arrested and the evidence seems thin, ask your attorney whether facial recognition was used. Police often don’t disclose it voluntarily.
Limit your photo exposure. Clearview scrapes public social media. Set your profiles to private. Remove old tagged photos. It won’t make you invisible, but it reduces the database’s accuracy.
Support bans. Multiple cities have banned government use of facial recognition: San Francisco, Boston, Minneapolis, Portland. Push for the same in your city.
Follow the legislation. The Surveillance Accountability Act (H.R. 8470) would require warrants for surveillance tools. Congress is considering it now.

Bottom Line

Angela Lipps spent five months in jail because a computer said she looked like someone. Police didn’t check if she was actually in the same state. They didn’t pull her bank records. They didn’t apologize when they were wrong. They still haven’t.

Thirteen known wrongful arrests from facial recognition. Five in 2025 alone. And those are just the ones we know about. How many people are sitting in jail right now because an algorithm made a match and a detective didn’t bother to verify it?

Facial recognition isn’t just inaccurate. It’s inaccurate in a way that destroys lives and then walks away without consequences. Until there are real penalties for departments that arrest people on AI matches alone, the 14th victim is already out there. They just don’t know it yet.

References

Published: April 28, 2026