Facial Recognition

The use of facial recognition surveillance technology has been shown to be inaccurate, racially biased, and a threat to personal privacy. In 2020 Detroit police officers arrested Robert Williams on his front lawn, in front of his wife and two young daughters, on charges that he had stolen watches from a Shinola store in Detroit. The arrest was based almost entirely on a facial recognition scan from security footage at the Shinola store, but it was dead wrong: Mr. Williams was not the man in the security footage and was nowhere near the store at the time of the theft. In 2021 the ACLU filed a federal lawsuit against the Detroit Police Department, alleging that the officers involved violated Mr. Williams’ rights under the Fourth Amendment and the Elliott-Larsen Civil Rights Act by arresting him on the basis of this flawed technology. Discovery in the case exposed systemic and comprehensive failures to use facial recognition technology responsibly and to train detectives in basic investigatory techniques and legal requirements. Meanwhile, in a second lawsuit brought by private counsel on behalf of another man falsely arrested under similar circumstances, we filed a friend-of-the-court brief in April 2023 highlighting the dangers of facial recognition technology and the failures we had uncovered during discovery. In June 2024 we entered into a settlement agreement, enforceable by the federal court for four years, that converts the Detroit Police Department from one of the nation’s worst abusers of facial recognition technology into a leader in imposing guardrails that limit possible abuses. Under the settlement, the police are prohibited from conducting lineups based solely on facial recognition leads, a procedure that creates a rigged lineup and lies behind each wrongful arrest using this technology in Detroit. The settlement also requires facial recognition examiners to disclose critical information about their searches (and the factors rendering them less reliable) to detectives, courts, and attorneys—information that can ensure each of those actors are aware of, and can act upon, exculpatory information arising from such searches. The settlement also requires training for police officers, including on how the technology misidentifies people of color at higher rates than other people. (Williams v. City of Detroit; Oliver v. City of Detroit; ACLU of Michigan Attorneys Phil Mayor, Ramis Wadood and Dan Korobkin, and legal interns Arshi Baig and Simon Roennecke; National ACLU Attorney Nathan Freed Wessler; Co-counsel Michael J. Steinberg of U-M Law School, with student attorneys Eilidh Jenness, Ben Mordechai-Strongin, Jeremy Shur, Deborah Won, Rihan Issa, Camelia Metwally, Seth Mayer, Jonathan Barnett, Lauren Yu, Will Ellis, Mickey Terlep, Brendan Jackson, Keenan McMurray, Julia Kahn, Lacie Melasi, Collin Christner, Ewurama Appiagyei-Dankah, and Nethra Raman.)

See further case background