This archive report was first published on 25 June 2020.
Published on June 25, 2020, a complaint filed in Detroit highlights the risks of facial recognition technology, which critics say can lead to racial bias and wrongful arrests.
Robert Williams, an African-American man, was wrongly arrested in January after a flawed facial recognition algorithm matched his face with surveillance footage from a robbery at a watch store.
Williams was held for 30 hours before learning he was wrongly identified, sparking concerns over the accuracy of facial recognition systems, particularly in distinguishing faces of black people.
"Though Robert Williams may be the first known case, he is likely not the first person who was wrongfully arrested and interrogated based off a bogus face recognition hit," the American Civil Liberties Union (ACLU) said in a statement.
The ACLU, representing Williams, is seeking a dismissal of the charges, an expungement of the arrest record, and a public apology to Williams.
As the use of facial recognition technology continues to rise, concerns over its accuracy and potential for racial bias have grown, with some firms, including IBM, Amazon, and Microsoft, announcing they would not sell the technology to police departments until regulations are in place.
However, many other systems are still widely used, and the ACLU is urging law enforcement agencies to stop using facial recognition technology as an investigatory tool.
