Georgia Man Arrested for a Crime In a State He Has Never Visited Based on Facial Recognition Tool: ‘Can’t Tell Black People Apart’

A Georgia man was jailed for over a week after authorities used botched facial recognition technology to wrongfully arrest him for a crime.

Baton Rouge Police Department detectives issued a warrant for Randal Reid, 28, in July for stealing luxury purses from a Louisiana consignment shop — using images from various mug shots, driver’s licenses and social media, his attorney told The Times-Picayune/The New Orleans Advocate. He was arrested in an Atlanta suburb in November 2022 as he drove to his mother’s home for Thanksgiving dinner.

“They told me I had a warrant out of Jefferson Parish. I said, ‘What is Jefferson Parish? ‘” Reid said. “I have never been to Louisiana a day in my life. Then they told me it was for theft. So not only have I not been to Louisiana, I also don’t steal.”

Reid is one of several Black people who have been arrested because facial recognition technology is “flawed and biased,” according to the ACLU. Research shows the technology has mistaken darker-skinned women for men about 31 percent of the time but has a 0.8 percent margin of error among white men. The algorithms work best identifying middle-aged white men but “can’t tell Black people apart,” the ACLU points out.

Several law enforcement agencies have shied away from the technology in response to the criticism. Authorities in New Orleans said they only use their system to generate leads, and detectives must go through a hierarchy before making a request through the Louisiana State Analytic and Fusion Exchange in Baton Rouge. The matches also must pass a peer review by other facial recognition investigators.

Patrick Grother of the National Institute of Standards and Technology told WIRED that facial technology error rates have dropped 90 percent since 2018. Still, Grother said, “there’s a considerable spectrum of accuracy” and “image quality remains an issue.”

Still, the best algorithms can be wrong more than 20 percent of the time, he added.

The Jefferson Parish Sheriff’s Office reportedly requested the algorithm adopted by Baton Rouge detectives, pegging Reid as one of three men who allegedly stole more than $10,000 in Chanel and Louis Vuitton purses over three days at the store in Metairie, Louisiana.

The ACLU says by using mugshots in facial recognition systems, law enforcement agencies are “supercharging” bias “with 21st-century surveillance technology” since it recycles data, and Black people are arrested at higher rates than whites.

Reid was booked into the DeKalb County Jail as a fugitive and released on Dec. 1, according to reports.

The man’s attorney, Tommy Calogero, said JPSO detectives “tacitly” admitted the error and rescinded the July warrant.

“I think they realized they went out on a limb making an arrest based on a face,” he said.

Back to top