The Detroit Police Department used facial recognition technology nine times in 2025 to identify suspects in violent crime investigations, a 91 percent decline from approximately 100 searches conducted in 2023, according to department data reported to the Board of Police Commissioners.
The sharp reduction follows a June 2024 settlement in Williams v. City of Detroit that imposed what civil rights advocates described as the strongest police department facial recognition policy in the United States. The case originated from the 2020 wrongful arrest of Robert Williams, who was detained at his Farmington Hills home after the department’s facial recognition system incorrectly flagged his driver’s license photo as matching a suspect in a Shinola store shoplifting case.
Under the settlement terms, negotiated by the ACLU and the University of Michigan Civil Rights Litigation Initiative, Detroit officers are prohibited from making arrests based solely on facial recognition results. Officers cannot conduct photo lineups derived directly from a facial recognition search without independent, corroborating evidence linking the identified individual to the crime under investigation. The department must also train all officers who access the technology on its limitations, including documented higher misidentification rates for people of color.
The settlement additionally required DPD to audit all cases since 2017 in which facial recognition was used to obtain an arrest warrant, creating a retrospective review of the technology’s role in past investigations. Department use of facial recognition had already begun declining in 2024, when it was deployed 28 times. Of the nine searches conducted last year, only one produced an actionable investigative lead, according to records reviewed by the Board of Police Commissioners.
The Williams case followed a separate wrongful arrest in 2023 involving Porcha Woodruff, who was eight months pregnant when Detroit officers arrested her based on a facial recognition match in a robbery and carjacking investigation. Both cases drew national attention to the risks of relying on automated facial matching in criminal investigations and prompted the department to revise its facial recognition policies multiple times before the formal settlement codified the current restrictions. The ACLU and the University of Michigan Civil Rights Litigation Initiative have described the resulting policy framework as a model for other jurisdictions considering guardrails on law enforcement use of biometric identification technology.
Sources: BridgeDetroit
–
By Ali Nassar-Smith
