A third person is suing the Detroit Police Department over a false arrest caused by the use of facial recognition technology.
Police arrested Porcha Woodruff, a 32-year-old Black woman, for robbery and carjacking based off a facial recognition hit from surveillance footage. The victim of the crime then picked her out of a photo line-up.
But Woodruff didn't do it, and charges against her were eventually dismissed for lack of evidence. Critics say this is more proof the technology is flawed, racially biased, and dangerous.
Tawana Petty, who heads the Detroit Digital Justice Coalition, called Detroit police “overreliance” on facial recognition during investigations “disheartening.”
“This was a totally, totally predictable circumstance that we've been shouting from the rooftops about for several years now,” Petty said.
Petty said it’s even more disheartening that when police arrested Woodruff, she was eight months pregnant. The crime had occurred only two weeks earlier, and the perpetrator was not visibly pregnant.
“You show up to a woman's door, you say ‘Two weeks ago you were involved in a carjacking’ and she answered the door eight months pregnant,” Petty said. “And you proceed? At the very least, there's an expectation of common sense.”
Phil Mayor, senior attorney with the ACLU of Michigan, said Woodruff’s case is part of a pattern of DPD investigations based on facial recognition that “line up a pipeline to make false arrests.”
Mayor noted that facial recognition hits are inevitably going to produce people who resemble the suspect. Police then put that image in a photo line-up, and eyewitnesses pick out the person who looks most like the perpetrator—the one provided by the facial recognition algorithm.
“It's a very suggestive lineup procedure that rigs the system by giving the person an image that necessarily is going to look something like the person who's accused of doing the crime,” Mayor said. “They are setting up a system that is custom-built to result in false arrests. And that's what it has done multiple times here in Detroit.”
Mayor noted that nationally, a total of six people are known to have been arrested based on faulty facial recognition matches. Three of those cases are from Detroit.
Detroit Police officially adopted the use of facial recognition software in 2019, following months of heated controversy. At the time, the department issued assurances that any facial recognition match would be used only as an investigative lead, and could never be used as the sole basis for an arrest. It also pointed to multiple measures in department policy meant to safeguard civil rights and prevent false arrests.
But Petty argues those safeguards clearly haven’t worked, and likely never will. “It's time that we remove this dangerous, racially biased technology from law enforcement hands, and we move on to something else,” she said.
In a statement, Detroit Police Chief James White called the allegations in Woodruff’s lawsuit “very concerning.” He added that DPD is “taking this matter very seriously,” but declined further comment, citing the need for additional investigation.
