A buyer drives forty minutes to a showing, walks into the kitchen, and realizes the bright, open room in the listing was partly invented. The lawn was greener online. The power lines out back were gone. That buyer is not making an offer, and they may be calling the state board. AI photo editing has made this easy, and 2026 is the year the rules started catching up.
What happened
California's AB 723 took effect January 1. According to NAR's REALTOR News, it requires agents to disclose digitally altered images that add, remove or change property elements, and to provide access to the original. It covers furniture, flooring, walls, landscaping, façades, floor plans and views. Routine lighting, color and cropping adjustments are excluded if they do not change how the property is represented. The same August article notes that New York is considering a similar bill, and that Minnesota's NorthstarMLS issued guidelines in July requiring AI-generated, AI-enhanced and virtually staged images to be clearly labeled and paired with an unaltered photo. NAR points agents to Article 12 of its Code of Ethics, which requires a "true picture" in advertising.
AI video is next. A HousingWire opinion column in June warned that tools turning still photos into simulated drone flyovers and walkthroughs can mislead buyers about how a property looks and how the footage was captured. It also noted Wisconsin's Act 69, which takes effect in 2027 and covers technology-altered advertising that creates a false or misleading impression.
Compliance is lagging. Inman's tech roundup on September 30 led with the finding that most AI-edited listing photos go undisclosed. And not everyone agrees new laws are the answer: in HousingWire's April reporting, an MLS executive argued existing rules just need enforcing, while an agent doubted enforcement can keep pace with AI.
Why it matters for your business
For agents and small brokerages, AI photo tools are tempting for good reasons. Virtual staging costs a fraction of physical staging. Decluttering, sky replacement and twilight shots used to require a pro editor. Now any agent can do them in a minute.
The problem is that the line between "enhanced" and "misleading" is exactly where liability lives. The rules are converging on a simple test: did the edit change what a reasonable buyer would believe about the property? Adding furniture to an empty room, labeled, is fine. Removing a crack in the foundation or a power line is not, labeled or not.
There is also a trust cost that never shows up in a lawsuit. A buyer who feels tricked by your photos does not trust your price, your disclosures or you. In a business that runs on referrals, that is expensive.
The counterargument is that buyers know photos are polished, and always have. Wide-angle lenses and good lighting have flattered homes for decades. True, but generative AI can now add and remove physical features, not just flatter them. That is a different thing, and regulators are treating it that way.
What we'd do
- Adopt the strictest rule you are likely to face. Even outside California, label every altered image and keep an unaltered version next to it. It is easier to follow one rule everywhere.
- Label specifically. "Virtually staged: furniture digitally added" tells a buyer more than "image enhanced."
- Never remove permanent features or defects. Power lines, cracks, stains, neighboring buildings and views stay as they are.
- Disclose simulated video. If a flyover or walkthrough was generated from stills, say so on the video itself.
- Keep originals and an edit log for every listing. If a complaint comes in, you want the receipts.
- Check your MLS and brokerage rules now. MLSs like NorthstarMLS are writing their own requirements, and they may apply even where state law is silent.
Our take: AI photo tools are good for agents and buyers when they help people picture a home, and bad for everyone when they hide what is really there. The agents who label clearly will look more trustworthy, not less. Disclose every time, and let the house be the house.
Want a system like this in your business?
A discovery conversation is where we map it: what to automate first, where the human checkpoints go, and what it would take.
Book a discovery conversation →