“Any Black Man Will Do”
A Transparency Framework for Eyewitness Identification in the Facial Recognition Technology Era
Ariam O. Kiflemariam
Abstract
Eyewitness misidentification has long plagued the criminal justice system. Historically and presently, Black and brown Americans are misidentified at rates multiple times higher than white Americans. The advent of facial recognition technology (FRT) has compounded these risks, introducing algorithmic biases that further entrench racial disparities while endangering constitutional protections for all. This article highlights the hidden dangers of police reliance on FRT in eyewitness identification, demonstrating how it exacerbates, rather than mitigates, wrongful arrests and convictions. Examining judicial responses in New Jersey, Washington, and Oregon, the article critiques existing standards that fail to account for the fallibility of human memory and the discriminatory flaws in both traditional identification procedures and AI-driven surveillance. It argues for a comprehensive transparency framework modeled after New York’s expansive open discovery law, which mandates the disclosure of identification procedures, equipping defense attorneys with the necessary tools to challenge unreliable evidence—including evidence flowing from FRT use, which this article argues falls within the scope of New York’s statutory scheme. By integrating scientific research on eyewitness fallibility with robust procedural safeguards, this framework seeks to balance the interests of defendants, governments, and the public, while ensuring that criminal prosecutions uphold constitutional protections rather than perpetuate systemic injustice.