When AI makes a mistake, it’s the agent who pays the price 

Though artificial intelligence is playing an increasing role in everyday life, guidelines and regulations concerning its use are still largely absent — and the situation is no different in the real estate industry.

"Agents punch in what they want, and then those outputs? They're running with it," real estate compliance consultant Summer Goralik explained during a recent episode of The REAL AI Guy podcast. "There's no gatekeeper. There's no review. There's no human approval."

This tendency has a lot to do with how busy real estate professionals are, the former California Department of Real Estate investigator said. But while agents are embracing this new tech to improve efficiency, it comes with risks.

"We still have to have humans behind the gate here watching this, preventing problems, correcting problems," she said. "If you're just using AI to get what you need and go, I fear that there's going to be a lot of misses — and there could be consumer harm when that happens."

Avoid 'playing with fire': The basic threshold for using chatbots and other AI tools should align with whatever an agent's local regulators say unlicensed assistants can and cannot do. "If you have an unlicensed person engaging with a consumer, you are playing with fire," Goralik said.

 "When you think about the chatbot engaging with the user who's a consumer, the second that chatbot starts to dole out advice, you should hear that buzzer in your ear because you're crossing the line. You are tempting fate. You're at risk — regulatory risk."

Have a policy and a plan for enforcement: Brokers should be knowledgeable about the AI tools their agents are using and should establish policy guidelines — but not just for the sake of saying they have them. "I help people create policy manuals, and then it will sit in a drawer," Goralik said. "They'll have their agent sign it. It sits in a drawer. It was expensive. They think that they checked off their box."

Simply having a policy isn't enough. Having an enforcement plan — and following through — is what will keep brokers and their agents on the right track, she added.

AI won't pay the price for its own mistakes: It's possible AI could continue to improve and eventually become better with compliance than humans are, Goralik suggested. But at the end of the day, the agent must remain involved in the decision-making process.

"What people don't realize is, if you have AI be your substitute unlicensed assistant, in my world, the unlicensed assistant is also in trouble. But AI is not going to get in trouble," she noted. If AI messes up, it'll be the agent whose license is at risk and who may have to afford legal fees. "You're the one that's going to pay the price: the human."

The regulators' perspective: Regulators want to protect consumers, but they're "not just out to get you," Goralik said. She has encouraged regulatory groups to speak "to their licensees, advise them, guide them at a minimum, put out an advisory" to help real estate professionals navigate AI use at a time when concrete rules haven't yet been written.

Regulators themselves may not be sure how to guide agents, she added, and some may prefer to wait for a consumer-submitted complaint before deciding how to act. "By doing an advisory ahead of time, it's almost like they're showing their cards — and maybe they don't have all of those cards in order yet."

Enforcement action likely: "There's going to be a misuse of AI because it's already happening," Goralik said. "The question is, where does it spill over into the regulator's realm?"

Looking ahead, "I do think there will be an enforcement action," she added. "And unfortunately, there might be some big firm that takes the hit on that."

Regulator action may be necessary for long-term clarity: "It's only going to take one complaint, one harmed consumer. And once that first enforcement comes out, that is going to be the best education that all of the brokers and agents can receive. It's going to show us exactly how the government is looking at this," Goralik said.

"It would be nice to have proactive, preventative advice," she added. "But I don't know if we're going to be that lucky."