What are the protected classes under the Fair Housing Act?

The federal Fair Housing Act (42 U.S.C. §3604) prohibits discrimination in housing on the basis of race, color, religion, national origin, sex, familial status, and disability. Following Bostock v. Clayton County (2020), HUD interprets “sex” to include sexual orientation and gender identity for enforcement purposes. Many states and cities add further protected classes, including source of income.

Does Fair Housing law apply to AI-generated listing content?

Yes. The law applies to the advertising itself, regardless of whether a person or an AI tool produced it. The agent and broker who publish the content are responsible for it. The FTC and HUD have both signaled increased scrutiny of automated and AI-driven systems in housing, which makes a compliance check on generated copy more important, not less.

Can I describe a listing as “family-friendly” or “perfect for a young couple”?

No. Phrases that target who should live in a home (“perfect for a young family”, “ideal for empty nesters”, “no kids”, “adults only”) implicate familial status, a protected class. Describe the property’s features instead - bedroom count, layout, yard - and let buyers decide whether it fits.

Is refusing a Section 8 voucher a Fair Housing violation?

Source of income is not a protected class under the federal Fair Housing Act, but it is protected by law in 21+ states and many cities. Language like “no Section 8” is unlawful in those jurisdictions and frequently overlaps with familial-status and race discrimination, so it is treated as high-risk and removed.

What should I write instead of “safe neighborhood” or “good schools”?

Replace subjective area claims like “safe neighborhood”, “good schools”, or “quiet community” with a concrete, verifiable fact - a low-traffic street, a gated entry, or a cited school rating (for example, “rated 9/10 on GreatSchools”) - rather than a demographic-sounding proxy. The same idea applies to ability framing: prefer “accessible” over outdated terms like “handicapped”, and describe the home rather than the resident.

Does ListingBooster guarantee my content is Fair Housing compliant?

No tool can guarantee that, and we do not claim to. ListingBooster runs a Fair Housing and financial-fidelity check inside the generation pipeline: the model is given hard compliance rules, generated copy is scanned for known violation patterns and steering proxies, and common violations are auto-rewritten before a draft reaches your approval queue. You still review and approve every post, and final responsibility for what you publish is yours. The goal is to make the safe version the default, not to remove your judgment.

Guide

Fair Housing for AI-Generated Content

AI can write a listing in seconds - including, if you let it, the phrases that draw Fair Housing complaints. This is a practical, sourced guide to what the law covers, the words that get flagged, and how to write them instead.

Last updated 2026-06-28. Educational only - not legal advice. Consult your broker or counsel for your jurisdiction.

Encrypted at rest
Stripe PCI Level 1
Fair Housing built-in
You own your content

The protected classes

The federal Fair Housing Act of 1968 (42 U.S.C. §3604) makes it unlawful to advertise housing with any preference, limitation, or discrimination based on a protected class. The seven federal classes are:

Race

Color

Religion

National origin

Sex

Familial status (children under 18)

Disability

The rule covers the advertisement itself, so it applies to copy a model wrote exactly as it applies to copy a person wrote. The agent and broker who publish remain responsible.

Sexual orientation and gender identity

In Bostock v. Clayton County(2020), the Supreme Court held that discrimination on the basis of sexual orientation or gender identity is discrimination “because of sex.” In 2021 HUD announced it would enforce the Fair Housing Act’s sex-discrimination provisions to include sexual orientation and gender identity. In practice, treat sexual orientation and gender identity as protected: never frame a listing around the orientation or identity of who “belongs” in a home or community.

Source of income

Source of income is not a federally protected class, but it is protected by law in 21+ states and a growing list of cities and counties. In those jurisdictions, “no Section 8” or “no vouchers” is unlawful. Because source-of-income exclusions also correlate with familial status and race, this language is high-risk even where it is not separately banned - leave it out.

Phrases that get flagged - and what to write instead

These are real patterns ListingBooster’s compliance scanner looks for. The fix is almost always the same: describe the property, not the people.

Familial status

“perfect for a young family”, “no kids”, “adults only”

Describe the home, not the buyer: “three bedrooms”, “fenced yard”, “single-level layout”.

Disability

“handicapped”, “able-bodied”

“accessible”; describe the home, not the resident, and omit ability framing.

Race / national origin

“exclusive neighborhood”, an ethnic/racial “community” label, “English-speaking”

Name a concrete feature; never characterize the people who live there.

Religion

“Christian community”, “near a church” as a selling point

“near places of worship” if location is genuinely relevant.

Source of income

“no Section 8”, “no vouchers”

Omit it. Refusing housing vouchers is illegal in many states and cities (see below).

Steering proxies

“safe neighborhood”, “good schools”, “quiet community”

Cite a verifiable fact: “low-traffic cul-de-sac”, “rated 9/10 on GreatSchools”.

The subtler risk: steering by combination

Individually innocent facts can become a steering signal when combined. A caption that stacks median income, a school rating, who is “moving in,” and a desirable-buyer profile can read as a demographic preference even though no single phrase is a banned word. This is exactly where AI-generated copy drifts, because a model optimizing for persuasive language will reach for aspirational, exclusivity-coded phrasing. A good compliance check looks at the whole caption, not just a keyword blocklist.

How ListingBooster handles this

Compliance is built into the generation pipeline, not bolted on at the end:

The model is given hard Fair Housing rules as part of every generation, so the first draft already avoids the common violations.
Generated copy is scanned against known violation patterns - the familial-status, disability, race, religion, source-of-income, and steering-proxy phrases above.
A semantic check looks at the whole caption for combination-steering, not just single banned words.
Common violations are auto-rewritten (“master bedroom” to “primary bedroom,” “handicapped” to “accessible”) before the draft reaches you.
Every post still requires your one-click approval before it can publish. The check makes the safe version the default; you stay the final reviewer.

No tool can guarantee compliance, and we do not claim to. The pipeline lowers the odds that a violation reaches your approval queue; reviewing and approving each post is still yours to do.

Frequently asked questions

Sources

Fair Housing Act of 1968, 42 U.S.C. §3604 (prohibited advertising and protected classes).
U.S. Department of Housing and Urban Development (HUD), Office of Fair Housing and Equal Opportunity.
Bostock v. Clayton County, 590 U.S. ___ (2020); HUD 2021 directive applying the sex-discrimination provisions to sexual orientation and gender identity.
National Association of REALTORS® Fair Housing guidance and Code of Ethics.
State and local source-of-income protections (21+ states and numerous cities/counties).

This guide is educational and not legal advice. Fair Housing law and state protections change; confirm the current requirements for your jurisdiction with your broker or counsel.

Let the safe version be the default

ListingBooster runs the compliance check on every generated post, then hands it to you to approve. Try it on a real listing with 25 free credits.

Start Free Compare ListingBooster