Crime, Law and Justice

In lawsuit, Minneapolis man says facial recognition tech led to his false arrest

A man looks past the camera
Kylese Perryman speaks to reporters at a news conference in Minneapolis on Wednesday.
Matt Sepic | MPR News

A Minneapolis man claims in a federal civil rights lawsuit filed Wednesday that authorities wrongfully jailed and charged him based on faulty facial recognition technology.

In September of 2021, Hennepin County prosecutors charged Kylese Perryman, 21, with a carjacking and the armed robbery of several women at a Mall of America parking ramp the following day.

Perryman's attorneys said that investigators using facial recognition technology falsely matched Perryman’s photo with security camera images of another young Black man allegedly seen shopping with the victims' credit cards at a Brooklyn Center Walmart soon after the robbery.

At news conference with Minnesota ACLU leaders, attorney Molly Jean Given said if prosecutors and Bloomington police had conducted even a basic investigation, they would have found overwhelming proof of Perryman’s innocence.

"While the suspects allegedly carjacked someone, Mr. Perryman was working the night shift at a Target warehouse. And like he did every time he worked that night shift, he clocked in and he clocked out. And he was clocked in when that car was stolen," Given said.

Given said Perryman was sleeping at the time of the mall robbery and was at a family member’s birthday party in Andover when the suspects took the victim’s credit cards to Walmart. She added that Perryman is significantly taller and thinner than the suspect in the security video and has different tattoos.

Perryman said he spent five days in jail before a judge released him with a GPS ankle monitor. He said it took months to clear his record with the help of a hired defense attorney.

“I feel like I had to prove my innocence more than they had to prove I was a suspect, and the system is not supposed to work that way,” Perryman said.

ACLU of Minnesota Legal Director Teresa Nelson said facial recognition technology is notorious for misidentifying people of color.

“Studies of facial recognition technology show that Black people are up to 100 times more likely to be misidentified than white men. That rate is even worse for Indigenous people,” Nelson said.

Spokespeople for Hennepin County and the city of Bloomington, which are both named as defendants, said they are unable to comment on pending litigation.

Volume Button
Now Listening To Livestream
MPR News logo
On Air
MPR News