The high-profile case of a Black man wrongly arrested this year wasn’t the first misidentification linked to controversial facial recognition technology used by Detroit police, the Free Press has learned.
Last year, a 25-year-old Detroit man was wrongly accused of a felony for supposedly reaching into a teacher’s vehicle, grabbing a cellphone and throwing it, cracking the screen and breaking the case.
Detroit police used facial recognition technology in that investigation, too.
It identified Michael Oliver as an investigative lead. After that hit, the teacher whose phone was snatched from his hands identified Oliver in a photo lineup as the person responsible.
Oliver was charged with a felony count of larceny in the May 2019 incident on West Warren Avenue in Detroit.
Oliver told his attorney he didn’t do it. Evidence in the case supported him.
Controversy over law enforcement using facial recognition technology is not new, nor is it confined to Detroit. But recent uprisings around the country in response to racial injustice in the wake of the death of George Floyd have again brought criticism of the technology to the forefront.
In Detroit, where police started using facial recognition software as an investigative tool in 2017, protesters have demanded the city stop using it, saying the error rate is high when used to identify people of color. City Council, which will consider extending a software contract to help pay for it, has been urged to vote no by some residents. Detroit’s civilian Board of Police Commissioners also has been discussing the department’s use of technology.
In the cellphone case, according to transcribed testimony, the teacher called 911 as he watched a group of students fighting. One student had a baseball bat and others were wrestling on the ground. The teacher used his cellphone to film the incident. The phone was recording when a young man reached into the teacher’s car and snatched the phone.
Oliver said the first thing that crossed his mind when his lawyer showed him the footage: “It wasn’t me.”
Oliver has tattoos up and down his arms. Those markings weren’t visible on the person captured on video. Oliver’s attorney, Patrick Nyenhuis, also noticed differences in the hairstyle and body type between the person in the video and his client, he said.
“It was obvious they had the wrong person.” Nyenhuis told the Free Press.
He took his concerns and pictures of his client to Wayne County assistant prosecutor Brian Surma, a supervisor in the office. Surma and the teacher reviewed photographs, determined Oliver was misidentified and both agreed the case should be dismissed immediately, a court transcript shows.
“We are convinced that there was a misidentification here,” Surma told a judge in September.
The case was tossed.
“I’m glad it’s all over,” Oliver told the Free Press this month.
Oliver, now 26, said he was nervous as his case proceeded last year because people still get convicted for crimes they didn’t commit. He questioned how his face ever got connected to the case.
During the investigation, police captured an image from the cellphone video, sent it for facial recognition and the photo came back to Oliver, the police report said.
After Oliver was singled out, a picture of his face was included in a photo lineup of possible suspects that was presented to the teacher.
A second person, a student, was also captured in the video with the suspect. The officer in charge of the case testified he didn’t interview that person though he’d been given that student’s name.
Police investigated Oliver’s case prior to a new policy governing the use of facial recognition software. It includes stricter rules on when Detroit police can use it. The technology is now used only as a tool to help solve violent felonies, Detroit police have said.
A spokesman for the department said Wednesday that he was looking into questions from the Free Press about Oliver’s case.
The Wayne County Prosecutor’s Office also now has more stringent protocol in place for facial recognition cases.
Evidence in Oliver’s case wasn’t reviewed by a supervisor in the prosecutor’s office prior to him being charged, spokeswoman Maria Miller said in an email. Current protocol requires a supervisor review all evidence in a facial recognition case prior to a charging decision. There also must be other evidence that corroborates the allegations in order to charge someone.
The prosecutor’s office is taking additional steps, Miller said. It will now be required that facial recognition cases be submitted to prosecutor Kym Worthy — the highest-ranking person in the office — for approval if an assistant prosecuting attorney and supervisor determine charges should be authorized.
Miller said the prosecutor’s office knows of no other cases in which people charged with crimes were misidentified, other than that of Oliver and Robert Williams.
Williams is a Farmington Hills man arrested in front of his family in January and accused of stealing high-end watches. Prosecutors and police have apologized for how that case was handled.
The case generated headlines across the country, including in the Free Press, the New York Times and the Washington Post. While in custody, Williams said he told police he wasn’t the man seen in a blurry image from store surveillance video.
“As a result of these two cases, we have a more stringent protocol in facial recognition cases,” Worthy said in an statement. “The cases will be reviewed during the warrant charging phase, prior to the preliminary examination, and again when the case is bound over to the Circuit Court in any case where facial recognition has been used as an investigative tool.”
She said she supports the use of the technology as an investigative tool only.
“In the summer of 2019, the Detroit Police Department asked me personally to adopt their Facial Recognition Policy,” she said. “I declined and cited studies regarding the unreliability of the software, especially as it relates to people of color.”
Studies have shown the technology, relying on computer algorithms, sometimes has trouble distinguishing human faces, especially with people of color.
Oliver and Williams are both Black.
Last week, Black Democrats in the Michigan House of Representatives called for a ban on the technology.
Detroit police chief James Craig, who is Black, has said he is a strong believer in facial recognition software. Last summer, he said police had used the technology about 500 times, then moved on to the next phase of investigation only 30% of the time.
Craig has blamed poor investigative work for what happened to Williams.
The Detroit police commission discussed Williams’ case during a meeting Thursday afternoon. Police gave a presentation and Craig said the situation should not have happened.
Williams also spoke and encouraged the technology to be banned, calling it racist.
Last month, board member Evette Griffie sought answers from police about Williams’ case, including a timeline of events and any discipline resulting from the misidentification.
“One mistake is too many,” she said at the time.
Griffie told the Free Press she wanted to know more about what happened.
“Do we have the proper checks and balances to make sure that what’s supposed to happen is actually happening?” she said.
Staff writers Nancy Kaffer and Paul Egan and the Associated Press contributed to this report.
Contact Elisha Anderson: [email protected]
This article originally appeared on Detroit Free Press: Detroit facial recognition technology has misidentified suspects