Tighter Policies Lead to Fewer Facial Recognition Searches for Detroit Police

Why It Matters

The sharp cutback highlights growing legal and public pressure on law‑enforcement biometric tools, reshaping market demand for facial‑recognition vendors. It also signals that unchecked use can trigger costly lawsuits and policy overhauls, affecting public‑safety budgets nationwide.

Key Takeaways

• Detroit police facial‑recognition searches fell 91% to nine in 2025.
• Only one of nine searches generated a lead, all targeting Black males.
• 2024 settlement stopped new spending; department now borrows MSP system free.
• Misidentifications sparked lawsuits, forcing stricter policies limited to violent crimes.
• Detroit schools use 1:1 facial ID, avoiding broader bias of 1:n matching.

Pulse Analysis

The Detroit Police Department’s retreat from facial‑recognition (FRT) reflects a broader shift as municipalities confront civil‑rights lawsuits and community backlash. After three high‑profile misidentifications—two involving Black men and a pregnant woman—settlements forced the department to rewrite its policy, restricting FRT to violent‑crime investigations and banning its use for protest surveillance. This regulatory tightening has not only curbed the number of searches but also eliminated new capital outlays; the department now relies on a cost‑free memorandum with Michigan State Police, illustrating how inter‑agency sharing can replace direct vendor contracts.

Industry analysts see Detroit’s experience as a cautionary tale for biometric vendors. The market, once buoyed by promises of city‑wide deployments, now faces a fragmented demand landscape where law‑enforcement agencies treat FRT as a last‑resort investigative aid rather than a routine tool. The 2024 settlement halted $2.37 million in annual spending, prompting vendors to pivot toward sectors with fewer civil‑rights constraints, such as private security and education. Yet even in schools, the technology’s application is carefully scoped—Detroit Public Schools employ a 1:1 ID‑matching system that sidesteps the bias‑prone 1:n searches used by police, highlighting a nuanced approach to risk management.

Looking ahead, the sustainability of FRT in policing hinges on two factors: robust governance and evolving AI capabilities. Advances in deep‑learning algorithms may improve accuracy across skin tones, but without transparent oversight, misidentifications will likely trigger fresh litigation, further eroding public trust. Policymakers at state and federal levels are already drafting legislation that could restrict or ban law‑enforcement use of biometric surveillance, echoing trends seen in Illinois and other jurisdictions. For agencies weighing FRT adoption, the Detroit case underscores the importance of aligning technology with clear, enforceable policies and community expectations to avoid costly legal fallout.