A black Michigan man falsely identified as a shoplifter by facial recognition software has filed a lawsuit seeking damages from the police.
Robert Williams, 43, is looking for an ‘undisclosed’ figure from the city of Detroit, his Police Chief James Craig and Detroit Police Detective Donald Bussa for his wrongful arrest in January 2020.
Williams, an auto worker, was arrested in front of his wife and young daughters on the lawn of his home in Farmington Hills, Detroit, after the photo of his Michigan driver’s license was linked to grainy surveillance footage of a shoplifting suspect.
‘This should never have happened, and I want to make sure this painful experience never happens to anyone else,’ said in a statement released by the American Civil Liberties Union (ACLU), which supports his case.
Williams’s case received national attention last June when the ACLU filed a complaint calling for an immediate cessation of use of the facial recognition software.
Several studies have shown that facial recognition systems are more likely to misidentify people of color than whites.
Robert Williams was arrested in January 2020 in front of his wife Melissa and children in his Detroit home after facial recognition software accidentally compared him to a suspected shoplifter.
Grainy surveillance footage of a shoplifting suspect taken in a Detroit watch store in 2018, left, was compared to Robert Williams’ state driver’s license, right
Williams’ ordeal began when facial recognition software purchased from DataWorks Plus by Michigan Police Department searched the driver’s license photos and compared it to grainy surveillance camera footage of an alleged thief stealing watches from a downtown Detroit store in 2018.
That led to Williams’ arrest in January 2020 for his wife and young daughters on their lawn in the Detroit suburb of Farmington Hills.
“I came home from work and was arrested in my driveway in front of my wife and daughters, who watched in tears because a computer made a mistake,” he said.
He was held for 30 hours in a “ filthy detention center in Detroit, where he had to sleep on an elevated cement slab due to overcrowding, ” the ACLU said.
The 75-page lawsuit filed in the US District Court for the Eastern District of Michigan reiterates that the use of facial recognition software should be discontinued.
“It is well documented that facial recognition technology is flawed and unreliable under the best of circumstances,” he says.
A “ match ” is not a likely cause and does not entitle the police to arrest, the lawsuit continued.
The technology is bad at accurately identifying black people, “especially in cases like those where the photo is grainy, the lighting is poor, and the suspect is not looking at the camera,” the ACLU said in an accompanying press release.
Detective Bussa, who was responsible for the arrest, had done “clearly sloppy, sloppy investigative work,” the lawsuit said.
“Bussa hasn’t even done a rudimentary investigation into Mr. Williams’ whereabouts during the shoplifting incident; if he had, he would have learned that Mr. Williams was driving home from work outside Detroit during the event in question and that he couldn’t be to blame, ”he said.
Police records show that the case began in October 2018 when five expensive watches went missing in the flagship store of Detroit luxury watchmaker Shinola.
An injury prevention officer later viewed the video that showed the suspect was a black man wearing a St. Louis Cardinals baseball cap.
“Video and photos have been sent to Crime Intel for facial recognition,” said a brief police report. Face recognition came back with a hit – for Williams.
At the top of the facial recognition report, prepared by the Michigan State Police, was a warning in bold, capitalized letters that the computer’s finding should be treated as an investigator, not a likely cause for arrest.
But Detroit detectives then showed a six-picture setup with Williams to the loss prevention officer, who positively identified Williams, according to the report.
The 43-year-old auto worker, pictured outside his home with his two daughters and wife Melissa, is demanding ‘undisclosed damages’ from the city of Detroit and the police
It took months for the police to issue an arrest warrant and several more before they called Williams at work and asked him to come to the police. It is not clear why.
Williams said he thought it was a joke. But they came to his house shortly afterwards, took him away in handcuffs and held him at night.
It was during his interrogation the next day that it became clear to him that he had been falsely identified by facial recognition software.
James Craig, the Detroit police chief also named in the suit, said The Detroit News facial recognition was not the cause of the wrongful arrest because it was “just bad detective work.”
He said Bussa had been demoted from commander to captain as a result of the botched arrest.
A day after Williams’ complaint in June 2020, Detroit Mayor Mike Duggan and Craig said his arrest report would be removed and his personal information removed from the police database.
Craig said facial recognition software was still in use, but with a lot more caution.
But the ACLU called for a complete cessation of its use.
Detroit police detained Williams for 30 hours in a “ filthy detention center in Detroit, where he had to sleep on a raised cement slab due to overcrowding. ”
“We know facial recognition technology threatens everyone’s privacy by making everyone a suspect,” said Phil Mayor, senior staff attorney for the Michigan ACLU.
“We have repeatedly urged the Detroit police to stop using this dangerous technology, but they insist on using it anyway. Justice requires that DPD and its officials are held accountable. ‘
The process is also supported by the University of Michigan Law School’s Civil Rights Litigation Initiative (CRLI).
Jeremy Shur, a student attorney at CRLI, said, “Cities across the country have banned police from using facial recognition technology for a reason.
“The technology is racially biased, flawed, and easily leads to false arrests of innocent people, just like our client.”
A 2019 study by the National Institute of Standards and Technology found that many facial recognition systems are more likely to misidentify people of color than whites.
It found that when conducting a particular type of database survey known as “ one-to-one ” matching, many facial recognition algorithms identified African-American and Asian faces 10 to 100 times more than Caucasian faces.
Hundreds of police departments quietly started using the software, despite the flaws.
But as awareness of its uses and its shortcomings has grown, many cities, including San Francisco, Minneapolis, New Orleans, and Boston, have banned its use.
Detroit activists have proposed reforms to end its use in the city.
But several bills introduced to ban or delay use of the software were not passed, the Detroit News reported.