Document Type

Article

Publication Date

12-2021

DOI

10.2139/ssrn.4101826

Abstract

Does law enforcement use of face recognition technology paired with eyewitness identifications increase the incidence of wrongful convictions in U.S. criminal law? This Article explores this critical question and posits that the answer may be yes. Facial recognition is frequently used by law enforcement agencies to help generate investigative leads that are then presented to eyewitnesses for positive identification. But erroneous eyewitness accounts are the number one cause of wrongful convictions, and the use of face recognition to generate investigative leads may create the conditions for erroneous eyewitness identifications to take place. This is because face recognition technology is designed to query a large database of faces to find look-alikes, and sometimes an innocent lookalike will resemble a suspect so closely that police may mistakenly select that person as an investigative lead, and an eyewitness may be unable to tell the difference between the lookalike and the actual suspect. This Article explores this possible problem and offers policy recommendations to help address it.

Publication Citation

William & Mary Bill of Rights Journal, Vol. 30, Issue 2, Symposium: Algorithms and the Bill of Rights, 337-372.

Share

COinS