by
Deposit photo
The Detroit Police Department has dramatically reduced its use of facial recognition to help identify violent crime suspects following a legal settlement in 2024 that limited the ways in which it can be used, BridgeDetroit reports.
Facial recognition software was used nine times last year, the department said, down 91 percent from 2023. In 2024, it was used 28 times.
At issue: Accuracy. Critics say the software has challenges when it comes to brown and Black faces, resulting in some false arrests.
“If it’s not being used (or) hardly at all, that’s a good thing. It’s something we really want to reserve for a last resort,” Board of Police Commissioners Member for District 7 Victoria Camille told BridgeDetroit.
Detroit police first purchased facial recognition software through South Carolina-based DataWorks Plus for $1 million in 2017. The department had no policy governing its use the first two years.
A lawsuit filed by the American Civil Liberties Union resulted in a settlement in 2024 that barred Detroit police from using the software to arrest people based solely on facial recognition results, or on the results of photo lineups directly following a facial recognition.
The case involved Robert Williams, who was at work in 2020 when he got a call from the Detroit police saying he needed to turn himself in. He had no idea why. When he got home, police were there and handcuffed him in front of his wife and two young children, according to an article in the Quadrangle, an online University of Michigan Law School publication.
He found out he was being arrested for felony larceny, the result of flawed facial recognition technology that wrongly identified him as the person who stole watches from an upscale store in Detroit.
Between his arrest and the settlement, the DPD falsely arrested two other Black individuals based on faulty use of facial recognition, according to the U-M article.
Following the lawsuit, DPD issued a statement:
"At the time of Mr. Williams' arrest, Department members were being guided by general policies governing technology, suspect identification, and arrest protocols. Following the incident, the Department created a policy specific to facial recognition that included three independent sign offs before being approved for use in an investigation.
"The policy also makes it clear that a facial recognition match can only be used as a tip to further an investigation, and that it cannot be used as the basis for someone being named as a suspect in an investigation."
