Orrich Lawson | Getty Images
Police in Louisiana reportedly relied on false facial recognition matches to secure a warrant to arrest a black man for a theft he didn’t commit.
Randal Reid, 28, has been in prison for almost a week after a false match led to his arrest, according to a report published Monday on the Times-Picayune/New Orleans Advocate newspaper website NOLA.com. . Reid told the newspaper he had never even been to Louisiana:
On November 25, Reid was stopped by local police on Interstate 20 in DeKalb County, Georgia, and left for a late Thanksgiving celebration with his mother.
“They said they had a warrant from the Jefferson Diocese. I said, ‘What is the Jefferson Diocese?'” Reed said. “I’ve never been to Louisiana in my life. Then they said it was for theft. Not only have I never been to Louisiana, I don’t steal.”
Reid was held as a fugitive in DeKalb County Jail, but was released on December 1, prison officials said.
Reid’s attorney, Tommy Calogero, said detectives at the Jefferson Parish Sheriff’s Office “implicitly” admitted the mistake and canceled the warrant. I think you’ve noticed,” Calogero said.
“Police could have checked his height and weight.”
Describing his time in prison, Reid said, “I wasn’t eating and I wasn’t sleeping. I’m thinking about these charges. I haven’t done anything. Correct ID.”
The case reportedly began in June 2022 with the theft of $10,000 worth of Chanel and Louis Vuitton purses in Metairie, Louisiana. Calogero said it was not difficult to confirm that Reid was not the culprit.
Calogero said the mole on Reed’s face was one of the factors that ultimately forced police to release him. Calogero also “estimated the difference between Reed and the purse thief he saw on surveillance footage at £40,” said the Times-Picayune/New Orleans Advocate report. Unlike Reed, the culprit had “flabby arms”.
“The police could have checked his height and weight, made an effort to talk to him, or asked him to walk around the house looking for evidence. He would have complied.” Calogero told the newspaper.
In this case, we don’t know exactly which facial recognition was used. In an earlier case, Jefferson Parish Sheriff Joe Lopinto’s office requested facial recognition analysis through his Rouge’s Louisiana Analytic and Fusion Exchange in Baton, which uses Clearview AI and his MorphoTrak system, the report said. said.
Clearview software compares your face to photos from social media and many other sources. “Using facial recognition technology, our platform contains over 30 billion facial image knowledge provided by public-only web sources, including news outlets, mugshot websites, public social media, and other open sources. It contains the largest database ever published,” says the company’s website. .
The sheriff’s office denied the newspaper’s warrant request
According to a Times-Picayune/New Orleans Advocate report, two warrants were issued for Reid’s arrest due to false facial recognition matches. The first was obtained by the Jefferson Parish Sheriff’s Office and the second was for another crime being investigated by the Baton Rouge Police Department.
“Detectives from the Baton Rouge Police Department are seeking to secure an arrest warrant for the JPSO’s lead claiming they were one of three men involved in the theft of another luxury wallet at a Jefferson Highway store that same week. adopted the identity, court records show,” the Times-Picayune/New Orleans Advocate reports.
According to the report, Lopinto’s office did not respond to a request for information and “issued an arrest warrant for Reid on July 18 and a formal request for a copy of any policies or purchases related to facial recognition, citing an ongoing investigation.” Another warrant obtained by Baton Rouge police “does not say how Lopinto’s office identified the lead,” the report said.
We will contact the Sheriff’s Department today and update this article as new information becomes available.
Privacy concerns and evidence of bias in facial recognition systems are accelerating moves to ban governments from using the technology. “Facial recognition software is particularly bad at recognizing African Americans and other minorities, women and young people, often misidentifying or failing to identify them, with a significant impact on certain groups.” , said the Electronic Frontier Foundation.
