Why home sellers should care about responsible AI use, and why their agents should care more

The conversation about AI in real estate has moved past "should we use it?" It is now about how to use it without losing clients' trust, your license, or both.

If you are an agent or broker, here is the uncomfortable part. The tools that make your business faster also create liabilities that many of your peers have not accounted for, and sellers are starting to notice.

The trust gap is real

Cotality's AI in Housing 2026 report surveyed recent and prospective homebuyers in the United States, Canada, the United Kingdom and Australia. Its findings should stop every listing agent mid-scroll.

Put simply, most buyers assume you are using AI, and very few of them trust it. That gap is your problem to manage, because if you do not, it becomes your seller's problem at the closing table.

For sellers, the stakes are concrete

A seller hires you to present the property honestly and attract serious, qualified buyers. AI can help you do that. It can also undercut the sale in three ways.

Altered listing photos erode buyer confidence

AI-enhanced and virtually staged photos have become a flashpoint. On November 13, 2025, the New York Department of State issued a trend alert to homebuyers and real estate professionals about AI-generated listing images, reminding licensees that advertising must give an honest and accurate depiction of the property. Agents around the country describe arriving at showings to find that an AI edit had added windows, staged furniture that could not fit the room, or turned a mud lot into a lawn.

When buyers feel misled, they do not just walk away from the house. They stop trusting the listing and the agent. That means fewer offers, more days on market and, in the end, less money for your seller.

AI-written descriptions can create liability

AI tools invent features, overstate amenities and get square footage wrong, and they do it in polished, confident language. The Fair Housing Act prohibits discriminatory advertising no matter who or what wrote the ad. If an AI draft calls a neighborhood "family-friendly" in a way that steers buyers, or mentions a "recently updated HVAC" that was replaced a decade ago, you have published it. The tool did not. You did.

Automated valuations carry bias risk

Your seller's price may be shaped by an algorithm before an appraiser ever walks through the door, and the research on fairness cuts both ways. Freddie Mac found in 2021 that appraisals came in below the contract price more often as the share of Black or Latino residents in a neighborhood rose. Some valuation vendors say their automated models reduce that gap. However, research published through HUD found larger valuation errors in majority-Black neighborhoods, and federal regulators now require lenders that use automated valuation models to have quality controls that include nondiscrimination. For sellers in communities that have been undervalued in the past, whether their agent's pricing tools are fair is a financial question, not an academic one.

The rules are changing quickly

• California. AB 723 took effect January 1, 2026. When a listing image has been digitally altered, beyond routine edits like lighting or cropping, licensees must include a reasonably conspicuous statement and give buyers access to the original. A willful violation of California's Real Estate Law can be prosecuted as a crime, and licensees also face discipline from the Department of Real Estate.
• Other states. Wisconsin has enacted a similar disclosure law that takes effect in 2027, and bills are pending in New York and New Jersey.
• The Code of Ethics. Articles 2 and 12 of the REALTOR® Code require honesty and a true picture in advertising, and Standard of Practice 12-10 prohibits misleading images. The Code does not spell out a labeling rule for altered photos, but a misleading edit can still violate it.
• MLS rules. Many MLSs now require virtually staged or altered images to be disclosed, including CRMLS, Bright MLS and MRED. Some go further. CRMLS, for example, requires the unaltered original to appear next to the altered image. Check your own MLS's current rules.
• Colorado. Colorado replaced its 2024 AI Act in May 2026 with a narrower law on automated decision-making technology that takes effect January 1, 2027. Housing remains one of the covered areas, so agents who use automated tools for decisions such as tenant screening should watch it closely.
• Fair housing penalties. The maximum federal civil penalty in an administrative fair housing case is $26,262 for a first violation and $131,308 for a respondent with two or more prior violations, before damages and attorney fees.

What responsible AI use looks like for a seller's agent

None of this is about avoiding AI. A 2025 Realtor.com survey found that 82% of Americans already use AI for real estate information. The technology is useful, and agents who use it responsibly will outperform those who do not.

Responsible use means four things.

1. Review every word before it goes public. Every listing description, social post and email. The confident tone is exactly what makes AI errors dangerous, because the output looks right even when it is wrong.
2. Disclose altered images everywhere they appear. Not only in the MLS, but on your website, social media, flyers and emails. Label them clearly and make the originals available. Even where the law does not yet require it, this is the direction the industry is moving.
3. Understand the tools behind your pricing advice. If an automated CMA or valuation model shapes your recommendation, know what data it uses, how it weighs comparable sales, and whether it has been tested for bias.
4. Tell clients where AI stops and your judgment starts. Buyers are not rejecting AI. They are asking for safeguards. Sellers deserve the same transparency, and offering it is a competitive advantage.

Make your commitment visible

When a client, a regulator or an insurer asks how you govern the AI tools you use every day, "we're careful" is not much of an answer. Written rules, training and a record of what you checked are.

For individual agents, NIRAP's RAICP™ (Responsible AI Certified Professional) credential covers AI literacy, risk identification, duty of care, documentation, fair housing obligations and client disclosure. The Realtor edition includes a Texas-approved continuing education course. For brokerages, NIRAP's RAICB™ certification verifies the firm's program as a whole.