Face recognition makes colored people shoplifters26. June 2020
Face recognition makes colored people shoplifters
New York, 26.6.2020
The ACLU has filed a complaint against the Detroit Police Department for the wrongful arrest of a man based on facial recognition error.
This is the first known case in the United States: Robert Julian-Borchak Williams, who is black, was arrested in January for one day after the Rank One facial recognition service falsely attributed him to a shoplifter.
The ACLU called on Detroit police to stop using facial recognition and described the technology as “faulty”. Rank One’s software linked Williams’ driver’s license photo to surveillance video of a shoplifter who stole five watches from a Shinola store in 2018.
When a picture of the shoplifter was shown, Williams told police, “Do you think all black people look alike?” The video shows that the officers released Williams after admitting that “the computer” must have made the wrong decision.
Police have yet to comment on the complaint. In a 2019 blog post, Rank One cited research confirming the accuracy of facial recognition systems and described the concerns as misunderstandings.
Even if the US police are stubborn, elsewhere they are reacting more responsibly:
Earlier this month, Amazon announced a one-year moratorium on the use of facial recognition systems by law enforcement, Microsoft suspended the use of facial recognition systems by law enforcement indefinitely until more federal regulations are in place, and IBM announced a permanent end to the supply, development, or research of general purpose face recognition.