State AI Laws Are Piling Up. A Plain-English Tracker for Agents

There is still no federal AI law for real estate, and none looks close. However, the states aren't waiting. As of early June, 78 chatbot-related bills were being debated in 27 state legislatures, and since then governors have been signing them into law.

You don't need to read every bill. Most of them aren't aimed at real estate. What helps is knowing which kinds of laws are being passed, and which ones land on you. Here is how I'd sort them.

1. Laws aimed at AI developers (mostly not your problem, but it reaches you)

Illinois signed the Artificial Intelligence Safety Measures Act on July 6. It requires the largest AI developers to publish how they assess catastrophic risk, report serious incidents, and, in a first for any state, submit to annual independent audits. It follows similar laws in California and New York. Most of the Illinois law takes effect January 1, 2027, with the audit requirement starting in 2028.

You're likely not a frontier AI developer, so none of this applies to you directly. It matters because your vendors depend on these companies. Expect vendor contracts to start referencing these frameworks. When you buy an AI tool, ask what risk framework the underlying model's developer publishes and what the vendor will tell you if something goes wrong.

2. Chatbot laws (this is the one to watch if you have a website assistant)

These are the biggest group by number. They cover disclosure that you're talking to an AI, protections for minors, companion chatbots, and in some bills, chatbots giving advice that would count as the unauthorized practice of a licensed profession if a person said it.

Hawaii signed a chatbot safety law in July that requires operator disclosures. New York's S7263 would go further and make operators liable for substantive professional answers, with no protection from an "I'm an AI" disclaimer. That one hasn't passed as of this writing.

If you have an AI assistant on your website or client portal, you may count as the operator, even if you didn't build the model. Read what it actually tells clients, and if it is explaining contract terms or recommending financing options, change that.

3. Deepfake laws (nearly every state has one)

Around 47 states now have some form of deepfake legislation. Hawaii's new law prohibits harmful digital imitations and requires disclosure of synthetic performers in advertising. At the federal level, the Take It Down Act requires covered platforms to remove deepfake content within 48 hours of a valid request.

For real estate this matters in two ways. First, your marketing: AI-generated people or voices in your ads may need a disclosure. Second, fraud: seller impersonation and cloned voices are the same technology. A law doesn't stop a scammer, so the callback habit still does the work.

4. Listing and advertising disclosure (the one aimed straight at the real estate industry)

California has required since January 1, 2026 that brokers and agents disclose when listing images are digitally altered and provide access to the originals. Lighting, cropping, and color correction are carved out as long as they don't change how the property really looks. Other states have bills moving through that target AI-altered listing photos and deceptive AI in housing ads.

This group is the most likely to become a routine part of the job, but it's the easiest to get ahead of. A short line in the listing, like "virtually staged," and a copy of the original photo in your file covers most of it.

5. Laws about automated decisions in housing and lending

Colorado passed the first broad state law on high-risk AI decisions in areas like housing, financial services, and employment. Its effective date was pushed back to June 30, 2026 and it has been the subject of continued changes since. On the lending side, a federal CFPB rule took effect July 21 pulling back on disparate impact theory in fair lending supervision, but Fair Housing Act liability and state fair lending laws still apply. Fannie Mae's Lender Letter LL-2026-04, which requires sellers and servicers to document how they use AI in origination and servicing, took effect August 6.

If you refer buyers to lenders or work with AI-driven screening tools, you don't need to become a fair lending expert. Just know which of your partners have a governance framework in place.

What to do about all this

• Keep a one-page list of AI tools your office uses and what each one does. You can't comply with rules you can't map.
• Write a short AI policy. Cover listing media, client-facing chatbots, and what can be put into a tool.
• Follow the strictest rule you're likely to face if you work across state lines. It's easier than managing a different standard in each one.
• Check your state real estate commission and MLS. They tend to act faster than legislatures and are more specific to your license.
• Revisit the list every quarter. The patchwork changes quickly, and a note saying you checked is part of your paper trail.

What to take from this

The rules are going to keep multiplying, and they won't line up. The habits underneath them are the same though: disclose what AI touched, know what your tools do, keep a human in the loop for anything that matters, and write down that you did. Get those in place and each new law becomes a small adjustment, not a scramble.

- Jason