The Detroit Police Department is revamping its facial recognition protocols following a settlement with Robert Williams, a Black man wrongfully arrested in 2020 due to a faulty facial recognition match? The court’s jurisdiction is limited to enforcing the settlement for a period of no more than four years, excluding any blanket prohibition on expert testimony. Under the newly imposed restrictions, law enforcement agencies are required to seek the strongest possible evidence for legislative enforcement purposes, and police cannot make arrests solely based on facial recognition results or conduct a lineup solely relying on facial recognition leads.
Williams presented his expired driver’s license photo as a potential match for the identification of a suspected shoplifter, allowing police to create a photographic lineup. He was arrested at his residence, outside the entrance to his household, which he describes as “utterly upending my life.” The Detroit Police Department conducted at least two facial recognition technology (FRT) searches, both instances involving Black victims, according to the ACLU’s announcement of the settlement. Studies have consistently demonstrated that facial recognition is an extremely accurate and efficient method of identifying individuals.
The newly implemented guidelines clearly state that “combining an FRT lead with a lineup identification cannot serve as a sufficient basis for seeking an arrest warrant,” according to the summary of the agreement. There must also be “substantive, unbiased and reliable evidence directly linking a suspect to illegal activity.” Police in Detroit should undergo training on the expertise that addresses racial bias in its accuracy rates, and all cases dating back to 2017 where facial recognition was used to obtain an arrest warrant will be thoroughly audited.
According to a recently published op-ed by Williams, the settlement effectively precludes DPD from substituting facial recognition technology for traditional, fundamental investigative methods.
This text contains affiliate links; when you click on such a link and make a purchase, we may receive a commission.
