The DRE Said Your AI Assistant is an Unlicensed Employee. (So... Why are Realtor® associations recommending them to be used for everything)
There's so much going on with artificial intelligence (AI) and real estate. AI is being sold to agents as both a panacea and manna from heaven: we don't have to do anything anymore but drop our contacts into the AI and wait for the clients to come breaking down our door to shower us with listings.
I've also seen, on Quora.com, questions from consumers about how their agents have offered to use AI on their listings but are requesting an additional 1% added to the listing commission. The listing agent gets 1% more than the buyer's agent because the listing agent is using AI.
“Let me call my assistant…”
Her name is Gemini
That AI assistant you are relying on to write listing descriptions, generate social media posts, build pre-listing packets, do timelines and every other thing? The one you’re talking about charging your client an additional 1% to use on their listing? The California Department of Real Estate has classified it as an unlicensed employee. And everything it touches falls under your broker's duty of supervision.
The NAR 2025 Technology Survey found that 68% of agents have tried AI tools, but only 20%use them daily. Another 22% use them weekly, 27% a few times a month, and nearly a third, 32%, have never used AI at all.
ChatGPT dominates the market at 58% usage among agents, followed by Google Gemini at 20% and Microsoft Copilot at 15%. Only 17% of agents report AI having a moderately positive impact on their business.
That gap between adoption and satisfaction is not a technology problem. Agents are being sold a tool without being told what it actually is or who is responsible when it gets things wrong.
Better call your broker.
Departments of Real Estate(s) throughout the United States are very clear on this. In March 2026, the California Department of Real Estate issued a formal Licensee Advisory on AI. The key sentence is worth memorizing: "A broker's supervisory obligation extends to the tools used to conduct licensed or unlicensed activities, including AI-powered software.
If an AI tool generates inaccurate information, misleading advertising, or improper communications with consumers, responsibility under current law rests with the licensee and their responsible broker, not the technology provider."
That same logic flows through state real estate commissions nationwide. The Alabama Real Estate Commission maintains a "Do's and Don'ts of Using AI" guide on its homepage, reminding licensees that "you, the licensee, are ultimately responsible for using it without violating license law."
The Texas Real Estate Commission's Legal Update II for 2026-2027 warns that using AI-generated content in advertising without verifying its accuracy "can lead to significant legal repercussions, including the propagation of false information that deceives the public." The tool does not assume the liability. The licensee and the broker do.
The California DRE went further in its Winter 2026 Real Estate Bulletin. It classified AI as one type of unlicensed assistant in the office, albeit a digital one.
“My assistant said…”
The DRE also warned that AI can talk an agent into trouble faster than a human assistant can. An AI assistant may talk an agent through an issue, but it can also quickly talk them into breaking regulatory requirements. The Bulletin called that a legal minefield, adding that any and all oversight falls squarely on the shoulders of the agent's employing broker.
“AI is a tool, but the agent and ergo their broker are in control at all times as the guiding hands assuring compliance with real estate law." Using AI tools to conduct licensed activity may be equivalent to asking an unlicensed assistant to do licensed activity, which is a violation of California real estate law.
Catch-22
I got a promotional email recently from a webinar sponsored by the California Association of Realtors Education. The subject line could have been cut and pasted from any AI vendor pitch.
It promised AI could run my newsletter, build pre-listing packets, write market data, and virtually stage a property. All in "a prompt or two." The hook was speed. The promise was leverage. More signs in more yards. Less time at a keyboard.
Then I read it a second time and saw what the email didn't say. State regulators look at that same list and see a compliance checklist from hell. The newsletter that hallucinates a crime statistic. The pre-listing packet built on fake comps. The virtual staging that hides a water stain behind a digital wall. Every line in that email is a liability if the agent using it doesn't understand what the regulators already figured out.
The marketing promise versus the compliance trap
Let me break down that promotional email line by line, because the gap between what it sells and what the law requires is where brokers get exposed.
The marketing says hand everything to ChatGPT. The compliance reality is you can never hand off final output. If an agent copies and pastes from ChatGPT into the MLS or a newsletter without a documented human verification process, the broker is legally exposed.
The DRE advisory states that "professional judgment must be exercised at all times, and AI outputs should be reviewed for accuracy and approved by the licensee before being relied upon in transactions or consumer communications."
For now
That promotional email made AI sound frictionless. Three prompts and you are done. The legal reality is the opposite. Every AI output is a document that needs review. Every AI tool is an unlicensed assistant that needs supervision. Every listing description generated by a machine is still your responsibility when it goes live.
The DRE warned that an AI assistant can quickly talk an agent into breaking regulatory requirements. That is a legal minefield. And any and all oversight falls squarely on the shoulders of the agent's employing broker.
