A Project of the University of Michigan Law School and the MDefenders Program

This motion argues that an unverified facial recognition alert cannot, without independent corroboration, establish probable cause for an arrest and that evidence obtained through such an arrest must be suppressed. Drawing on expert testimony and scientific research, including a 2024 National Academies report on facial recognition technology, the motion explains how image quality, database composition, algorithmic error rates, demographic disparities, and the absence of trained human review undermine the reliability of automated facial recognition identifications. The motion emphasizes that facial recognition results should be treated as investigative leads rather than positive identifications, particularly when the underlying images, algorithms, and error-rate data have not been preserved or disclosed.

The motion further argues that the exclusionary rule applies notwithstanding the existence of an outstanding arrest warrant because officers relied on unverified technology, failed to follow departmental safeguards, and used excessive force without confirming the individual’s identity. It challenges the application of the independent source, inevitable discovery, and attenuation exceptions, emphasizing that reckless or flagrant police misconduct warrants suppression to deter future constitutional violations.

File Type: docx
File Size: 106 KB
Categories: 4th Amendment, Evidence, Expert Testimony, Facial Recognition, Forensics, Identifications, Police, Race
Author: Louisiana, National