Racial Segregation and the Data-Driven Society: How Our Failure to Reckon with Root Causes Perpetuates Separate and Unequal Realities

Abstract

This Essay asserts that in the United States racial segregation has and continues to play a central evolutionary role in the inequalities we see reproduced and amplified by data-driven technologies and applications. Racial segregation distorts and constrains how data-driven technologies are developed and implemented, how problems of algorithmic bias are conceptualized, and what interventions or solutions are deemed appropriate and pursued. After detailing the foundational aspects of how racial segregation has evolved over time and its less obvious social, political and epistemic implications for White Americans, the demographic group that dominates the technology sector, this Essay explores how racial segregation affects algorithmic design, analysis and outcomes. It concludes with analysis of how prevailing approaches to evaluating and mitigating algorithmic bias are insufficient and why a transformative justice framework is necessary to adequately examine and redress algorithmic bias as well as improve the development of data-driven technologies and applications. This Essay illustrates how critical analysis of racial segregation can deepen our understanding of algorithmic bias, improve evaluations of data-driven technologies for social and racial equity concerns, and broaden our imaginations about what meaningful redress of technology-mediated harms and injustices should include.