In 2018, Robert Williams was wrongly accused of theft in Detroit. His driver’s license photo was mislabeled as possibly matching the man seen on security video at a Shinola watch store. After a lawsuit, the city of Detroit has agreed to pay him $300,000 in settlement.
As part of the agreement, Detroit will change the way it uses facial recognition technology to solve crimes. The police will no longer arrest people based on facial recognition results alone or a composite of photos generated by a facial recognition search. Instead, they must conduct further investigation to determine if there is evidence to support the identification made by the technology.
Robert Williams expressed his excitement about these new safeguards but hopes that they will not lead to the use of facial recognition technology altogether. The lawsuit settlement was announced by the American Civil Liberties Union and the University of Michigan Law School Civil Rights Initiative, which argued that facial recognition technology is flawed and racially biased.
Detroit Police Chief James White announced these new technology-related policies last August amid ongoing disputes over their use. These policies stated that there must be evidence other than facial recognition for police to believe that a suspect has the means, ability, and opportunity to commit a crime. As part of the settlement, Detroit police will review cases from 2017 to 2023 in which facial recognition was used. Prosecutors will be notified if arrests are made without independent evidence.
According to ACLU attorney Phil Major, individuals arrested based on facial recognition scans and lineup results often face pressure to plead guilty. This pressure is especially strong for people with criminal records, resulting in longer sentences and increased suspicion from both law enforcement and prosecutors.
In conclusion, this lawsuit settlement highlights the need for better safeguards when using facial recognition technology in law enforcement investigations. It also brings attention to potential biases in this technology and calls for more research into its effectiveness and limitations.