An AI rental listing takes seconds to generate and reads well. That is the risk. A draft that sounds polished can still describe the renter you supposedly want, promise a dishwasher the unit does not have, or sit next to a photo with the power lines edited out. Once you post it, the words and images are yours.
This guide is about AI tools specifically: why AI drafts introduce steering language, what happened to HUD's 2024 guidance on digital advertising and ad delivery, how to check generated amenity claims, the new California rule on AI-altered listing photos and a review checklist to run before anything goes live. It is part of our property marketing guides for independent landlords.
Note: This article is general education, not legal advice. Fair housing and advertising rules differ by state and city, and rules on AI and edited images are changing quickly. Facts are as of November 2026. Consult a local attorney if you are unsure whether a listing or ad campaign complies.
Why AI drafts can introduce steering language
AI writing tools produce text by predicting what typically comes next, based on the huge amount of existing writing they learned from. Real estate marketing is full of copy that describes people instead of property: "perfect for a young couple," "ideal for professionals," "great family neighborhood," "quiet, mature building." Ask a tool to make your listing appealing and it can reach for the same patterns, because they are common, not because they are lawful.
Steering language shows up in AI drafts in a few predictable ways:
- Ideal-renter lines. Closing sentences that describe who the unit "suits" or is "perfect for." These are the most common problem, and they often appear even when your prompt says nothing about renters.
- Neighborhood character. Phrases about who lives nearby, how "safe" or "family-oriented" an area is, or what kind of "community" it has. The tool fills gaps with impressions it cannot verify.
- Echoing your prompt. If you type "target audience: grad students" or "looking for a quiet older tenant," the tool will write that preference into the ad.
- Lifestyle framing. "Bring your kids to the park next door" or "no stairs to slow you down" turn a feature into a statement about who belongs.
The legal test does not care how the words got there. Under 42 U.S.C. 3604(c), it is unlawful to make, print or publish, or cause to be made, printed or published, an ad that indicates a preference, limitation or discrimination based on race, color, religion, sex, handicap, familial status or national origin. HUD's regulation at 24 CFR 100.75 lists words, phrases, photographs and symbols among the ways an ad can do that. Courts read ads as an ordinary reader would, so a line you did not write and did not notice counts the same as one you chose.
For the specific words and phrases to watch for, and safer property-based alternatives, see the separate guide in this series on fair housing advertising words to avoid. The rest of this article focuses on the parts that are unique to AI.
How to prompt an AI tool so the draft starts clean
The best fix is upstream: give the tool facts, not a target renter. A short, structured prompt produces a draft that needs far less editing.
- Start from a fact sheet. List bedrooms, bathrooms, square footage, rent, deposit, fees, utilities, lease length, pet and smoking rules, parking, floor level, stairs or elevator, appliances, recent updates and distances to named places such as a bus stop or grocery store.
- Tell the tool what not to do. Add a line such as: "Describe only the property, its features, its terms and distances to named places. Do not describe the ideal renter, residents, neighbors, schools' quality or neighborhood safety. Do not add any feature not in the list."
- Leave out people. Do not include a target audience, a description of current or past renters or anything about the neighborhood's residents.
- Ask for structure. Request a headline, two opening lines, a bulleted features list and a terms block. Structured output is easier to check line by line against your fact sheet.
Even a careful prompt is not a guarantee. Models do not always follow instructions, which is why the review step below is not optional. Our guide on how small landlords can use AI tools safely covers the broader ground rules for AI in a rental business, including privacy and keeping a human in every decision.
Accuracy: checking generated amenity claims
AI tools fill gaps. If your fact sheet does not say whether the unit has central air, a draft may say it does, because many listings like yours mention it. If you mention a nearby park, the tool may add a walking time it made up. These are not rare glitches; generating plausible detail is how the tools work.
Inaccurate amenity claims cause practical and policy problems. Renters who arrive expecting in-unit laundry leave annoyed, and the listing site may act: Zillow's Rental Listing Quality Policy requires listings to be kept current and accurate and prohibits fraudulent or misleading information, and lets Zillow remove listings at its discretion.
Check every generated claim against the fact sheet, and pay extra attention to these categories, where AI drafts most often overreach:
| Claim type | What AI tends to add | How to verify |
|---|---|---|
| Appliances and laundry | Dishwasher, in-unit washer and dryer, "stainless" finishes | Walk the unit or check your property record |
| Utilities and fees | "Utilities included," "no fees" | Compare to the lease and your fee schedule |
| Size and layout | Square footage, "open concept," room dimensions | Use measured figures only |
| Pets and parking | "Pet friendly," "off-street parking" | Match the exact written policy and costs |
| Distances and walk times | "Steps from," "5 minutes to downtown" | Measure with a map tool; state miles or blocks |
| Schools and neighborhood | School names, ratings, "safe," "vibrant community" | Delete opinions; keep only named, checkable places |
| Renovations | "Newly renovated," "brand-new kitchen" | State the actual work and year |
A related risk is the AI that searches the web on a renter's behalf. Renters increasingly ask chat assistants to find apartments, and those tools summarize your listing text. Accurate, factual listings are summarized correctly; vague or inflated ones are not. Our guide to how renters use AI to find rentals covers what that means for your listing.
AI-altered photos and California's new disclosure rule
AI photo tools can virtually stage an empty room, replace a gray sky, brighten a dark hallway or erase a neighbor's fence. Some of these edits are routine. Others change what the renter will actually find.
California has drawn a line. In 2025 it passed AB 723, which requires real estate licensees who use digitally altered images to advertise a home for sale to disclose the alteration and give access to the original. AB 2025, which Governor Newsom's office lists as signed on September 27, 2026, brings the same approach to rentals. According to the Assembly Privacy and Consumer Protection Committee's analysis of the bill:
- A rental ad that uses a digitally altered image must include a statement that the image was altered, plus a link, web address or QR code where renters can see the original, unaltered image.
- An image counts as digitally altered if photo editing software or artificial intelligence was used to add, remove or change elements such as fixtures, furniture, appliances, flooring, walls, paint color, landscaping, the facade or floor plans, or things outside the property such as utility poles or neighboring buildings.
- Routine adjustments such as lighting, sharpening, white balance, color correction, cropping or exposure do not count, as long as they do not change how the property is represented.
New California laws passed in a regular session generally take effect on January 1 of the following year, so AB 2025 is expected to apply from January 1, 2027. Confirm the start date and final wording in the chaptered text before relying on it, since the committee analysis describes an earlier version of the bill.
Outside California, there may be no specific photo disclosure law, but the same accuracy standards apply. A photo that hides a busy road or shows furniture the unit does not include can mislead renters and run afoul of listing site policies. A practical rule for every state: if an edit changes what a renter would see in person, label it and keep the original available, or do not use it. Our property photography guide covers how to get strong photos without heavy editing.
HUD's 2024 guidance on digital ads and algorithmic delivery
Listing text is only half of the AI picture. Ad platforms also use automated systems to decide who sees an ad, and that can create fair housing risk even when the ad itself is clean.
On April 29, 2024, HUD's Office of Fair Housing and Equal Opportunity issued guidance on applying the Fair Housing Act to advertising through digital platforms. It warned that audience selection tools, lookalike audiences and algorithmic ad delivery can steer housing ads away from protected groups, sometimes without the advertiser's direction or knowledge. It suggested that advertisers flag housing ads as housing when a platform asks, review the source of any custom audience lists and monitor campaign results.
HUD withdrew that guidance on September 17, 2025, along with several other guidance documents, in a memo stating that the withdrawal is not necessarily final. The withdrawal changed HUD's guidance, not the law. Section 3604(c) still applies, and 24 CFR 100.75 still lists selecting media or locations for ads that deny particular segments of the housing market information about available housing as a discriminatory practice.
For a small landlord, the practical steps are the same either way:
- Declare housing ads as housing. Meta, for example, treats housing as a special ad category. Its developer documentation says advertisers cannot choose a specific gender, ages are generally fixed at 18 to 65 and older, ZIP code targeting and location exclusion are not supported, U.S. locations must cover at least a 15 mile radius, and lookalike audiences are unavailable.
- Do not build custom audiences from your past renters. A list of current or former renters can reproduce their demographics in whoever the platform shows the ad to.
- Keep the targeting broad and the records clear. Save the settings you used for each campaign alongside the ad itself.
AI also shows up after the inquiry, in screening and approval tools. That raises different rules; see our guide to AI tenant screening and fair housing rules.
A review checklist before you post an AI rental listing
Run this checklist on every AI-generated listing, even if you used the same prompt last month. It takes about ten minutes.
- Read the last two sentences first. That is where ideal-renter lines usually appear. Delete any sentence about who the unit is "perfect for," "ideal for" or "great for."
- Scan for people words. Flag every word about age, family, gender, faith, origin, health, ability or a type of person, and rewrite it as a fact about the unit or remove it.
- Strip neighborhood opinions. Remove "safe," "family-oriented," "up-and-coming," "exclusive" and similar impressions. Keep named places and distances.
- Match every amenity to the fact sheet. If a feature is not on your list, confirm it in the unit or delete it.
- Check the numbers. Rent, deposit, fees, square footage and dates must match your records exactly.
- Check local protected classes. Remove anything about vouchers, income type or other classes your state or city protects.
- Review every photo. Label any image that adds, removes or changes elements of the property, and keep the original available. In California, follow AB 2025's disclosure and link requirements once they apply.
- Check the ad settings. If you are paying to promote the listing, choose the housing category and avoid narrow or custom audiences.
- Save the final version. Keep the published text, photos, run dates and targeting settings together.
For how a clean, accurate description fits into the rest of your listing, see our rental listing optimization playbook, and for the wider compliance picture, our fair housing compliance guide.
Frequently asked questions
Is it legal to use AI to write a rental listing?
Yes, nothing in federal fair housing law bans using AI to draft a listing. But the Fair Housing Act applies to whoever publishes the ad, so you are responsible for every word and photo an AI tool produces once you post it.
Why does AI add discriminatory language to rental listings?
AI writing tools learn from large amounts of existing marketing copy, which often describes an ideal renter or characterizes a neighborhood. Asked to make a listing appealing, a tool can reproduce phrases like "perfect for young professionals" that describe people rather than the property.
Do I have to disclose AI-edited photos in a rental listing?
In California, AB 2025, signed in September 2026, requires rental ads that use digitally altered images to say so and to point renters to the original. Elsewhere, edited photos that change what the unit looks like can still be misleading, so label them or use unedited photos.
Is HUD's 2024 guidance on digital housing ads still in effect?
No. HUD withdrew its April 2024 guidance on advertising through digital platforms on September 17, 2025. The Fair Housing Act and HUD's advertising regulation at 24 CFR 100.75 still apply to how ads are written and placed.
What to do next
AI drafting saves time only if the review is fast too, and review is fast only when the facts are already written down. Landlords who keep a current record of each unit's features, terms and policies can check an AI draft in minutes; those who rely on memory end up rewriting it or, worse, posting it unchecked.
Shuk has a live AI assist, powered by OpenAI, that generates property listing descriptions from the details you enter. It produces a draft for you to review and edit, not a finished ad, and Shuk does not claim it checks the draft for fair housing compliance, so the checklist above still applies. The details it works from are captured when you add a property and its units, including beds, baths and square feet and direct yes or no answers on pets and smoking. Year-Round Marketing keeps the listing visible on Shuk's own listing site and syndicates it to Zillow. Shuk does not edit or label listing photos, so photo disclosures remain your responsibility.
At as low as $2 per unit per month, with no setup fees and no contract, and with White Glove Onboarding included at no additional cost, Shuk makes fast, reviewed, accurate rental listings feasible for landlords and property managers up to 300 units.
Book a demo at shukrentals.com/book-a-demo to see how listing description generation, property records and Year-Round Marketing work together so an AI draft becomes a listing you can stand behind.
