Algorithmic Bias and the Modern Interpretation of the Fair Housing Act
By: Jonathan Curbelo Edited By: Hannah Becker and Christina Ding
When the Fair Housing Act (FHA) was signed into law in 1968, it was intended to eliminate the visible architectures of segregation: restrictive covenants, explicit redlining maps, and obvious refusals to rent due to race. Decades later, the contest over housing equity has shifted from the landlord’s office to the automated server. Today, tenant screening and lending for mortgages is moderated by complex proprietary algorithms. While these tools promise objective efficiency, they often act as boxes that perpetuate historical inequities under the guise of mathematical neutrality. As artificial intelligence (AI) saturates the real estate market, existing laws must be revisited to meet the demands of digital redlining in the twenty-first century.
Legally, it is challenging to enforce fair housing standards against algorithmic models because the distinction between disparate treatment and disparate impact remains highly contentious among the judiciary. In accordance with the 2015 Supreme Court ruling in Texas Department of Housing and Community Affairs v. Inclusive Communities Project, Inc., the Court upheld that the FHA covers disparate impact claims, practices that are not intentionally discriminatory but have a disproportionately adverse effect on protected groups. [1] However, under Executive Order 14179, issued in December 2025, the national policy approach for AI emphasizes a minimally burdensome approach to regulation, thus complicating the issue of civil rights claims and technological innovation. [2]
To date, 42 U.S. Code § 3604 serves as the primary federal check against housing discrimination, declaring it unlawful to "make unavailable or deny, a dwelling to any person because of race, color, religion, sex, familial status, or national origin." [3] In the context of AI, the Department of Housing and Urban Development (HUD) has struggled to clarify whether software developers can be held liable for the discriminatory outputs of their models. In 2024, HUD issued guidance stating that the FHA applies to AI systems just as it does to human actors, yet the specific standards for proving bias in a hidden algorithm remain unclear in federal appellate courts. [4]
Beyond disparate impact, the nature of the technology also complicates discovery efforts in court. Because many lending models are protected as trade secrets, plaintiffs often cannot access the specific variables, such as zip codes or shopping habits, that act as proxies for race. The U.S. Copyright Office and patent laws currently offer robust protections for these machine processes, making it difficult for regulators to audit them for creative or discriminatory inputs. [5] In response, states like California and Colorado have moved faster than the federal government, passing legislation that requires companies to conduct impact assessments for high-stakes automated decision systems. California’s recent AB-331, for instance, mandates that deployers of automated tools provide a plain-language explanation of how AI influenced a housing or credit denial. [6]
Federally, the closest legislation to a unified AI housing standard is the "Discriminatory Effects Standard" reinstated by HUD in 2023. This standard requires a three-step burden-shifting test. A plaintiff first must prove that a practice causes a discriminatory effect. Second, the defendant must prove that the practice is "necessary to achieve one or more substantial, legitimate, nondiscriminatory interests." Finally, the plaintiff must show how those interests could be served by a less discriminatory alternative. [1] However, with the emergence of generative credit modeling, where algorithms create their own scoring criteria, defendants can argue that the interest of the model is too complex for a human being to grasp, and, therefore, a less discriminatory alternative does not exist.
The impact of AI in the housing market has outpaced the slow-moving pace of the FHA and Lanham Act’s existing consumer protections. As current litigation and the contrast between FTC authority to regulate against unfair practices and HUD's authority to combat disparate effects show, there is an immediate need for a single federal AI housing standard. That need is underlined by the necessity of striking a balance between technological progress and Fourteenth Amendment promises of equal protection under law. A comprehensive federal framework that incorporates algorithmic auditing, clarity in data proxies, and clear liability for the developers of discriminatory software is necessary.
NOTES:
U.S. Department of Housing and Urban Development. 2023. “Reinstatement of HUD’s Discriminatory Effects Standard.” Federal Register. March 31, 2023. https://www.federalregister.gov/documents/2023/03/31/2023-05837/reinstatement-of-huds-discriminatory-effects-standard.
Trump, Donald J. 2025. “Ensuring a National Policy Framework for Artificial Intelligence.” The White House. December 11, 2025. https://www.whitehouse.gov/presidential-actions/2025/12/eliminating-state-law-obstruction-of-national-artificial-intelligence-policy/.
“42 U.S. Code § 3604 - Discrimination in the Sale or Rental of Housing and Other Prohibited Practices.” n.d. LII / Legal Information Institute. https://www.law.cornell.edu/uscode/text/42/3604.
U.S. Department of Housing and Urban Development. 2024. “Guidance on the Application of the Fair Housing Act to the Use of Artificial Intelligence and Algorithms.” April 29, 2024. https://www.hud.gov/sites/dfiles/FHEO/documents/AI_Guidance_2024.pdf.
United States Copyright Office. 2024. “Copyright and Artificial Intelligence.” July 2024. https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-1-Digital-Replicas-Report.pdf.
Bauer-Kahan, Rebecca. 2025. “California Assembly Bill 331 - Automated Decision Tools.” California Legislative Information. 2025. https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260AB331.
BIBLIOGRAPHY:
Bauer-Kahan, Rebecca. “California Assembly Bill 331 - Automated Decision Tools.” California Legislative Information, 2025. https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260AB331.
LII / Legal Information Institute. “42 U.S. Code § 3604 - Discrimination in the Sale or Rental of Housing and Other Prohibited Practices,” n.d. https://www.law.cornell.edu/uscode/text/42/3604.
Trump, Donald J. “Ensuring a National Policy Framework for Artificial Intelligence.” The White House, December 11, 2025. https://www.whitehouse.gov/presidential-actions/2025/12/eliminating-state-law-obstruction-of-national-artificial-intelligence-policy/.
U.S. Department of Housing and Urban Development. “Guidance on the Application of the Fair Housing Act to the Use of Artificial Intelligence and Algorithms,” April 29, 2024. https://www.hud.gov/sites/dfiles/FHEO/documents/AI_Guidance_2024.pdf.
U.S. Department of Housing and Urban Development. “Reinstatement of HUD’s Discriminatory Effects Standard.” Federal Register, March 31, 2023. https://www.federalregister.gov/documents/2023/03/31/2023-05837/reinstatement-of-huds-discriminatory-effects-standard.
United States Copyright Office. “Copyright and Artificial Intelligence,” July 2024. https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-1-Digital-Replicas-Report.pdf.