What are the AI real estate disclosure laws agents need to follow in 2026?

Why AI Disclosure Laws in Real Estate Exist Now

The wave of state-level AI disclosure rules hitting real estate in 2025 and 2026 did not appear out of nowhere. They are a direct response to a specific behavior pattern that became impossible to ignore: agents and listing platforms using generative AI to alter listing photos, stage empty rooms, swap faces in listing videos, and even synthesize voiceover narration, then publishing the result as if it were an unedited representation of the property. The National Association of REALTORS® has framed this practice as a form of "catfishing" buyers, and consumer-advocacy groups have pushed for statutory fixes rather than relying on voluntary ethics guidance. The result is a patchwork of state laws, federal guidance, and platform-level rules that now govern when and how AI-generated or AI-altered real estate media must be labeled.

Also worth reading: What are the MLS virtual staging disclosure rules and compliance standards for agents? · How do AI real estate valuation models work in 2026, and what role does virtual staging play in accuracy? · What are the rules and regulations surrounding AI real estate listing compliance?

The single most important thing to understand is that the disclosure obligation is triggered by the act of materially altering a representation of the property, not by the tool used to do it. Whether you used Photoshop in 2009 or a diffusion model in 2026, if the image, video, or audio a buyer sees does not faithfully represent the actual condition, layout, or contents of the home, you are now in the disclosure zone in several jurisdictions. California led the way by codifying this principle for listing photos, and New York followed with a broader synthetic-performer disclosure rule that captures AI-generated hosts in listing videos. Other states are actively drafting similar statutes, and the Transparency Coalition's June 26, 2026 legislative update tracks at least a dozen active bills.

For agents and brokerages, the practical consequence is that "we used AI" is no longer a marketing brag — it is a compliance trigger. The same disclosure that protects consumers also protects the listing agent from misrepresentation claims, fair-housing complaints, and class-action exposure when a buyer claims they relied on an altered image to make an offer.

The Core Disclosure Rules You Must Follow

California's law, which took effect and has been enforced through 2025 and 2026, requires that any AI-altered listing photo carry a clear and conspicuous disclosure that the image has been modified. The disclosure must be visible in the listing itself, not buried in a terms-of-service link or a fine-print footer. Consumer Reports and the Consumers Union backed the underlying California bill specifically because voluntary disclosure had failed — buyers were routinely shown virtually staged rooms, digitally added furniture, and even AI-swapped landscaping without any indication that the photo did not match the property.

New York's synthetic performer disclosure law, analyzed in detail by McDermott Will & Schulte, is broader in scope. It requires disclosure whenever a synthetic performer — including an AI-generated avatar, a deepfake of a real person, or a digitally cloned voice — appears in advertising content. For real estate, this captures AI listing videos where a synthetic host walks through a property, narrates a tour, or appears in a thumbnail. The law does not require that the AI be disclosed as "AI" specifically; it requires that the audience be informed that the performer is not a real human being, or not the human being they appear to be.

At the federal level, the Federal Trade Commission has signaled through enforcement actions and guidance that materially misleading AI-generated imagery in commercial contexts — including real estate — can be treated as a deceptive practice under Section 5 of the FTC Act. While there is no single federal "AI real estate disclosure statute," the FTC's stance means that even in states without a specific AI disclosure law, an agent who uses AI to materially mislead a buyer can face federal consequences. HousingWire's reporting on the "disclosure test" agents now need to apply is built around this FTC overlay.

What Counts as "AI-Altered" Under These Laws

The legal definitions vary, but the operational definition that most brokerages have converged on is: any digital modification that changes what a buyer would see if they physically walked through the property. This includes virtual staging (adding furniture to an empty room), virtual renovation (showing a kitchen with new cabinets that do not exist), AI sky replacement, AI lawn enhancement, decluttering (digitally removing a seller's furniture), and AI-generated listing photos of rooms that do not exist in the home. It also includes AI-generated listing videos, AI voiceover narration, and AI-generated virtual tours.

What generally does not trigger disclosure is basic color correction, white balance adjustment, lens correction, HDR processing, and standard perspective correction — the kind of edits that have been part of real estate photography for two decades. The line is drawn at changes that affect what the buyer perceives about the property itself, not changes that affect how the camera captured it. This distinction matters because it tells you where the legal risk actually sits: in the content of the image, not in the post-processing pipeline.

A useful mental model is the "walk-through test." If a buyer walked through the door with their phone camera, would they see what is in the listing photo? If the answer is no — because the room is empty, the furniture is different, the wall color is different, or the view out the window is different — the image has been materially altered and disclosure is required in California and in any state that has adopted a similar rule.

How to Disclose Properly (and Improperly)

The most common mistake agents make is treating disclosure as a one-time checkbox. The California law and the New York synthetic performer rule both require that disclosure be "clear and conspicuous," which in practice means visible at the point of viewing, not hidden behind a link. A disclosure that says "some images have been virtually staged" in 6-point gray text at the bottom of a listing page does not meet the standard. A disclosure that appears as a visible badge or label on each altered image — for example, "Virtually Staged" or "AI-Altered Image" overlaid on the photo itself — is closer to compliant.

For listing videos, the disclosure must appear before the viewer has been misled. A disclosure that appears in the final three seconds of a 90-second video, after the synthetic host has already walked the buyer through the home, is too late. The disclosure should appear at the opening of the video, in the thumbnail, or as a persistent on-screen label during the synthetic performer's appearance. McDermott Will & Schulte's analysis of the New York rule emphasizes that the disclosure must be "effectively communicated," which courts have interpreted to mean communicated before the consumer could form a material impression based on the synthetic content.

For audio — AI-generated voiceover, AI-cloned agent narration, or synthetic testimonials — the disclosure should be spoken at the beginning of the audio segment, not just written in a transcript or description. The same principle applies: the audience must know what they are hearing before they form an impression.

Comparison of Major Disclosure Regimes

The table makes clear that the regimes overlap but are not identical. An agent who complies with California photo rules may still violate the New York synthetic performer rule if they use an AI avatar in a listing video. An agent who complies with both state rules may still face FTC action if the overall impression of the listing is materially misleading. And an agent who complies with all three may still have their listing removed by the MLS or platform if it violates the platform's content policy.

Practical Steps for Agents and Brokerages

The first step is a listing-media audit. Pull every active listing and identify which photos, videos, and audio assets have been AI-altered or AI-generated. Tag each one in your listing system so that disclosure can be applied consistently. This is not a one-time project — new listings come on every week, and the audit must be ongoing.

The second step is to update your listing templates and your MLS input forms to include disclosure fields. Most MLS systems have added "virtual staging" or "AI-altered media" fields in 2025 and 2026 in response to the state laws. Use them. Do not rely on free-text descriptions, because they are inconsistent and hard to enforce across a brokerage.

The third step is to train every agent and listing coordinator on the walk-through test. Anyone who touches a listing photo or video should be able to answer the question: "If a buyer walked through this property, would they see what is in this image?" If the answer is no, the image needs a disclosure label before it goes live.

The fourth step is to document your disclosure practices. Keep a record of which images were altered, what tool was used, what the disclosure label says, and when it was applied. If a buyer later files a complaint or a regulator opens an investigation, this documentation is your first line of defense. The absence of documentation is treated by regulators as evidence that the disclosure was an afterthought, not a policy.

The fifth step is to review your vendor contracts. If you use a third-party virtual staging company, an AI photo enhancement service, or a synthetic video provider, your contract should require them to flag every asset they deliver as either "unaltered," "virtually staged," or "AI-generated." The disclosure obligation does not transfer to the vendor — it stays with the listing agent and the brokerage — but a vendor who cannot tell you which assets are altered is a vendor you should replace.

Common Mistakes That Trigger Enforcement

The most common mistake is the "subtle enhancement" — using AI to brighten a room, add a sunset to the sky outside the window, or sharpen the lawn, and then publishing the result without disclosure because the change feels minor. Under the California walk-through test, the change is material if the buyer would not see the same thing in person. A brightened sky is not material; an AI-generated sunset over a view that does not exist is material. The line is not about how big the change looks — it is about whether the change reflects reality.

The second most common mistake is the "AI disclaimer in the footer" — a single line at the bottom of the listing page that says "some images may be virtually staged." This does not meet the clear-and-conspicuous standard in California, and it does not meet the effectively-communicated standard in New York. Disclosure must be tied to the specific asset, not buried in a generic page-level statement.

The third most common mistake is the "synthetic host without disclosure" — using an AI avatar to present a listing video, often because the real agent does not want to be on camera, and failing to disclose that the host is not a real person. The New York rule was written precisely to capture this case, and the McDermott Will & Schulte analysis makes clear that the FTC has also signaled enforcement interest.

The fourth most common mistake is the "stale disclosure" — adding a disclosure when the listing first goes live, then editing the photos later (for example, swapping in an AI-staged version after the property has been on the market for 60 days) without updating the disclosure. The disclosure must reflect the current state of the media, not the state at listing launch.

When to Act and What It Costs

The disclosure laws are already in effect in California and New York, and enforcement has begun. The California Attorney General's office has issued guidance, and local associations of REALTORS® have circulated compliance checklists. The New York rule is being enforced by the Attorney General's office as well, with civil penalties that can reach several thousand dollars per violation. The FTC has not announced a specific AI-real-estate settlement as of mid-2026, but its enforcement posture under Section 5 is well established.

The cost of compliance is low compared to the cost of non-compliance. Adding a disclosure label to each altered image is a few minutes of work per listing, and most MLS platforms now provide the field for free. Training an agent on the walk-through test takes an hour. Updating vendor contracts takes a single negotiation cycle. By contrast, a single misrepresentation complaint can cost a brokerage tens of thousands of dollars in legal fees, settlement, and license-discipline proceedings — and a class action can cost orders of magnitude more.

For brokerages that have not yet implemented a disclosure program, the right time to act was last quarter. The second-best time is now. The third-best time is never, because the regulatory trajectory is clear: more states, more platforms, and more enforcement.

How AI Virtual Staging Fits Into This Landscape

AI virtual staging is one of the most common triggers for the disclosure obligation, and it is also one of the most defensible uses of the technology when disclosed properly. A virtually staged photo of an empty living room, clearly labeled as virtually staged, helps a buyer visualize the space without misleading them about the property's actual condition. The disclosure does not diminish the value of the staging — it preserves it, by ensuring the buyer understands what they are seeing.

The mistake is to treat the disclosure as a marketing liability. In practice, buyers respond positively to transparently staged listings, because the disclosure signals that the agent is being honest about the property. The listings that generate complaints are the ones where the staging is presented as reality — where the buyer shows up to an empty room and feels deceived. Disclosure is not just a legal requirement; it is a trust-building mechanism that protects the agent's reputation and the brokerage's brand.

For agents who use AI virtual staging tools, the operational rule is simple: every staged image gets a visible label, every staged video gets an opening disclosure, and every staged audio asset gets a spoken disclosure. The label should be specific ("Virtually Staged" is better than "AI-Enhanced"), visible (overlaid on the image or in the listing metadata), and consistent across the brokerage. Tools that automate this labeling are now standard in the major virtual staging platforms, and agents should prefer tools that include the disclosure by default rather than tools that leave it as an opt-in afterthought.

The Bottom Line

AI real estate disclosure laws in 2026 are not a speculative future — they are an active compliance regime in California, New York, and a growing list of other states, overlaid by FTC enforcement and platform content policies. The disclosure obligation is triggered by material alteration of property representations, not by the use of AI itself. The compliance cost is low, the enforcement risk is real, and the reputational upside of transparent disclosure is significant. Agents and brokerages that treat disclosure as a default practice — not a checkbox — will be the ones that thrive as the regulatory landscape tightens.