As lenders, servicers and vendors race to invest in AI and are touting their AI prowess alongside countless companies across the nation's economy, Brian Woodring, chief information officer at Newrez issued a warning.

"I don't think you could look at the industry and not think we're in the middle of an AI bubble," said Woodring. 

The executive evoked the dot-com bubble of the 2000s, but balanced that view by mentioning how tech giants emerged at the same time. Others speaking Tuesday suggested they're bullish on AI but emphasized that firms need to identify strong use cases and tangible results. 

Experts also echoed frequent industry forecasts regarding AI enhancing, or replacing workers. Terry Schmidt, CEO and director of Guild Mortgage, said employees will become more efficient at what they do and that human communication skills will always be in demand. Woodring gave the example of a chatbot's potential weaknesses in trying to solve some customer problems. 

"When [borrowers are] struggling to make their payments, who do they want to talk to?" said Woodring. "So I think of irreplaceable things like empathy and trust, the ability to listen and the ability to communicate as being effectively irreplaceable."

The Newrez leader also predicted the cost to service and originate loans, which has cut into lenders' profits in recent quarters, would decrease in the next five years. Schmidt, whose company uses its own proprietary tech, cautioned firms who are sitting on the sidelines. 

"I think it's going to be difficult for companies that don't get on the bandwagon quickly to continue to be competitive, I think it's going to be harder," she said. 

Industry attorneys warn of risks in AI adoption

The trade group featured numerous AI sessions and standing room-only crowds participated in exercises such as "vibe coding," in which non-engineers created loan calculators, marketing templates and other mortgage website functions with a literal click of a button. 

Attorneys Tuesday afternoon urged companies to implement strong risk management as state and federal lawmakers mull new legislation. Seven of the nation's most populous states are proposing, have proposed, or are studying laws around AI and bias mitigation. 

Companies budgeting for AI often aren't including testing, said Wendy Lee, managing partner at the LOGS Legal Group. Firms can begin by testing AI tools against themselves, by asking the platforms to show the source of their output.  

Brian K. Stucky, CEO of AI governance firm Decision-X and senior advisor to the MBA's Mortgage Industry Standards Maintenance Organization, called on companies to evaluate how vendors are using AI. He shared an anecdote of a vendor which could not define what a protected class was when asked about the data it handles.

"I think all the vendor contracts you've had probably need to be tossed out and rewritten for AI," he said. 

Even if a vendor doesn't promote itself as an artificial intelligence solution, it's likely using the tools, said Anna Craft, a partner at the Bradley law firm. She alluded to cybersecurity risks and recalled the now long-standing guidance by firms for employees not to plug sensitive data into tools like ChatGPT.

"Because they have your data, and if they have something that results in your data escaping in the wild, then you are just as liable as they are," she said.