AI in the Racial Wealth Gap: Deciding our Future (Full report)
Published June 2026
Racial Wealth Gap: The persistent disparity in accumulated assets between racial groups. As of 2022, the median White household holds ~$285K vs. ~$45K for the median Black household — a $240K+ gap widening since the 1960s.
Applicant Tracking System (ATS): Software that screens job applications by scanning for exact keyword matches, filtering out candidates before any human review
Automated Valuation Model (AVM): AI tools used to estimate home values. When trained on historically biased data, AVMs undervalue homes in communities of color.
Bossware: Digital surveillance + automated decision systems used to monitor workers. Intensifies labor insecurity and disproportionately affects workers of color in gig, warehouse, and service sectors.
Cascading Algorithmic Harm: When AI-driven denial in one domain (e.g., employment) triggers harm in others (savings depletion → housing denial → credit damage) because shared infrastructure connects all three.
Digital Redlining: When algorithmic systems in lending/credit systematically charge communities of color higher rates or restrict access — even without race as an explicit variable.
FICO Score: Dominant credit scoring model built on historical data from an era of explicit exclusion. Penalizes “credit invisibility” — a condition disproportionately affecting Black, Latino, and Indigenous borrowers.
Shared Credit Infrastructure: Credit bureaus (Equifax, TransUnion, Credit Karma) and background check databases that feed into housing, employment, and lending simultaneously — meaning a single negative marker cascades across all domains.
Transparency Deficit: The gap between AI’s role in decisions and people’s ability to know or contest them. Participants consistently reported not knowing whether AI was involved or how to appeal.