AI in Residential Real Estate: Efficiency Gains and Equity Gaps

Abstract

Artificial intelligence is moving through residential real estate faster than the academic evidence needed to evaluate it. We organize the literature around a simple tension: AI can make prediction, screening, matching, pricing, and generation more efficient, but it also changes who gets access, who captures surplus, who bears errors, and who faces risk. Using an equityefficiency framework, we classify roughly 145 papers across six residential domains and four AI mechanism categories. The evidence map reveals a sharp asymmetry: twenty-two of twentyfour cells lack peer-reviewed evidence on equity outcomes, even though industry deploys AI in eighteen cells and federal regulation is active in nine. Existing research is concentrated in automated valuation and mortgage credit; rental pricing, leasing automation, tenant screening, property management, and agentic workflows remain lightly studied. We translate these gaps into a research agenda anchored by regulatory shocks, vendor deployments, and residential REIT operating disclosures.

Keywords: Residential Real Estate, Machine Learning, Algorithmic Fairness, Automated Valuation Models, Algorithmic Pricing, Fair Housing, Large Language Models, Agentic Systems