AI Regulation Tracker / Regulation in force

HUD Loosened the AI Housing Rules. Your Disparate-Impact Liability Did Not

HUD moved to rescind its disparate-impact regulations in 2026, and it is tempting to read that as relief for AI tenant-screening tools. It is not. The Fair Housing Act and the liability it creates for biased outcomes are still here. Here is the gap landlords and proptech are walking into.

By Anthony Guerriero · Reviewed by The Leveraged Years Editorial Desk · Published June 28, 2026 · Last updated June 28, 2026

HUD Loosened the AI Housing Rules. Your Disparate-Impact Liability Did Not

The Leveraged Years AI Regulation Tracker

In January 2026 HUD moved to rescind its disparate-impact regulation under the Fair Housing Act, but the Act itself, and the disparate-impact liability recognized under it by the Supreme Court in Inclusive Communities (2015), remains in force. An AI tenant-screening tool that screens out protected groups at higher rates can still create liability, regardless of intent, so landlords who read the rollback as a green light are exposed. Primary sources: the Fair Housing Act and HUD's rescission notice.

Regulatory Briefing

Bindingness: Binding Law (statute remains; regulation rescinded)
Scope: United States / Sector: housing providers, property managers, proptech, and tenant-screening vendors
What Changed: In January 2026 HUD announced plans to rescind its disparate-impact regulations under the Fair Housing Act, part of a broader rollback. The regulation is the procedural framework. The Fair Housing Act itself, and the disparate-impact liability the Supreme Court recognized under it, remain in force, including for AI tenant-screening tools that produce discriminatory outcomes.
Who This Affects: Landlords, property managers, and proptech companies that use AI or algorithmic tools to screen tenants or target housing advertising, and the counsel advising them.
Current Status: In Force (Fair Housing Act); disparate-impact regulation being rescinded
Effective Date: HUD rescission announced January 2026; Fair Housing Act liability ongoing
Primary Source: The Fair Housing Act (42 U.S.C. 3601 et seq.) and HUD's 2026 action to rescind its disparate-impact regulation, via HUD: https://www.hud.gov/program_offices/fair_housing_equal_opp
Secondary Analysis: Nixon Peabody, "HUD rescinds disparate impact regulations" (January 21, 2026): https://www.nixonpeabody.com/insights/alerts/2026/01/21/hud-rescinds-disparate-impact-regulations
Workflow Impact: A housing provider that treats the rescinded regulation as permission to relax its AI screening oversight is reading the wrong document. The statute still prohibits practices that produce a discriminatory effect, and an AI tool that screens out protected groups at higher rates is exactly such a practice. The defense is the same as it was: know what your tool does, test it, and be able to justify it.
Operational Next Step: Inventory every AI or algorithmic tool used to screen tenants or target housing ads, ask the vendor what data it weighs and whether it has been tested for disparate impact, and keep your own record of that testing, because the regulation going away does not move the liability off your desk.
Last Verified: June 28, 2026

A rollback that changes less than it looks like

In January 2026 HUD moved to rescind its disparate-impact regulations under the Fair Housing Act, part of a wider pullback on fair-housing enforcement. For a landlord or a proptech founder, the headline reads like relief. The rule that made disparate-impact claims easier to bring is going away.

Read the second sentence before you relax. The regulation HUD is rescinding is a procedural framework. The Fair Housing Act is a statute, and the disparate-impact theory of liability under it was recognized by the Supreme Court in Texas Department of Housing and Community Affairs v. Inclusive Communities Project (2015). A regulation can be rescinded by an agency. The statute and the case law cannot, and they are what create the exposure when an AI screening tool produces a discriminatory result.

Why AI tenant screening is the sharp edge

Tenant screening is where this lands hardest, because it is the part of housing that quietly went algorithmic. Thousands of tools now score applicants on data of varying quality, and a model that weighs eviction records, credit data, or criminal history can screen out protected groups at higher rates without anyone intending it to. That higher rate is the definition of disparate impact, and it does not require intent.

HUD itself flagged this in 2024 guidance making clear the Fair Housing Act applies to tenant screening and housing advertising that rely on algorithms and AI. The enforcement posture has shifted since, but the legal theory the guidance described is still good law. Coverage of AI leasing tools has tracked a string of federal cases testing exactly this exposure, alongside a surge in related accessibility claims.

The gap landlords are walking into

The danger is a timing mismatch. The regulatory signal is loosening at the same moment the technology is spreading, which invites housing providers to add AI screening with less oversight precisely when they have more of it to oversee. The provider who reads the rescission as a green light and turns off the testing is the one most exposed, because the underlying liability did not move.

This is the lesson the briefing's title carried: the federal rules loosened, your liability did not. A rescinded regulation changes who is pushing enforcement. It does not change what the statute prohibits or who pays when a tool discriminates.

What to do now

The work is unglamorous and it is the whole defense. Inventory the AI and algorithmic tools you use to screen tenants or target housing ads. Ask each vendor, in writing, what data the tool weighs and whether it has been tested for disparate impact across protected classes. Run or obtain that testing, and keep your own record of it. Treat a tool you cannot explain as a liability you cannot defend.

None of that depends on HUD's regulation. It depends on the statute, which is still here, and on your ability to show that the tool you used did not produce a discriminatory effect. The providers who keep testing through the rollback are the ones who will not be surprised when the enforcement winds shift again.

Frequently Asked Questions

Did HUD eliminate disparate-impact liability for AI in housing?

No. In January 2026 HUD moved to rescind its disparate-impact regulation, which is a procedural framework, but the Fair Housing Act and the disparate-impact theory of liability recognized under it by the Supreme Court in Inclusive Communities (2015) remain in force. An AI tenant-screening tool that produces discriminatory outcomes can still create liability.

Does the Fair Housing Act apply to AI tenant screening?

Yes. HUD's 2024 guidance made clear the Fair Housing Act applies to tenant screening and housing advertising that rely on algorithms and AI, and that legal theory remains valid. A tool that screens out protected groups at higher rates can be a discriminatory practice regardless of intent.

What should a landlord or proptech company do now?

Inventory every AI or algorithmic screening and ad-targeting tool, ask vendors what data each weighs and whether it has been tested for disparate impact, obtain or run that testing, and keep records. The rescinded regulation does not move the statutory liability off the housing provider.

Is disparate impact still the legal standard if there was no intent to discriminate?

Yes. Disparate-impact liability turns on discriminatory effect, not intent. A facially neutral AI tool that disproportionately screens out a protected class can create liability under the Fair Housing Act even with no intent to discriminate.