Color-Blind Biometrics? Facial Recognition and Arrest Rates of African-Americans in Maryland and the United States
Year: 2019
Type: dissertation
Abstract: This thesis provides the first empirical analysis of statewide police facial recognition systems to determine if these systems affect the arrest share of African-Americans in Maryland and other implementing states.The question stems from the research of those working at the intersection of the privacy, civil liberties, and technology spaces, who in recent years have expressed concerns about bias in the algorithms underlying rapidly proliferating police facial recognition capabilities.Using arrest and crime data from the Federal Bureau of Investigation's Uniform Crime Reports, combined with data from the Census Bureau's American Community Survey, I create a panel dataset spanning the period 2006-2016, and then run the data through a series of regression models.I find inconclusive results when measuring the effect of facial recognition systems nationally, as well as when examining Maryland's alone.This lack of clarity suggests a need for a moratorium on police use of facial recognition until a better understanding of the effects of the technology is achieved, and the public debate regarding its ethics matures.
Author: Matthew Aaron Schwartz
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3100434493