“Fannie and Freddie have a big advantage,” Seguin said. “They look at the loan four months after it’s closed. It’s easy to see, ‘Hey, the mailing address has changed,’ and pull reports and find out that it’s been rented and listed for rent on NMLS. The lender doesn’t have that information up front, and could have done everything right, and they are still going to lose if the borrower is untruthful.”
Glad to see action taken
Seguin said that fraud typically occurs one of two ways in the mortgage industry. It is either fraud for profit, usually driven by industry insiders, or fraud for housing, when the borrower typically alters the numbers to make sure they qualify.
“You have fraud for profit and fraud for housing, where it’s probably the old 80-20 rule,” he said. “No one has an exact estimate, but fraud for housing is probably 80% of the fraud, but only 20% of the losses. Where the fraud for profit is only 20% of the time, but it’s probably 80% of the losses. And that’s where you start seeing those trends that really start adding up to big dollar losses for the lender.”
This is where the use of AI technology can help identify trends that could stop fraudulent transactions before they are completed, rather than months later.
Fannie Mae partners with Palantir to fight mortgage fraud using AI, starting with multifamily loans. Bill Pulte highlights the potential to reduce costs and risk, while Priscilla Almodovar emphasizes early fraud detection to protect future loans.https://t.co/dQxDhA5mUm
— Mortgage Professional America Magazine (@MPAMagazineUS) May 29, 2025
“It’s harder for an originator to identify one loan at a time,” Seguin said. “You need some tool to come in and say, ‘That loan officer had 20 loans in the last year. They’re all self-employed borrowers. All of them have used the same CPA.’ What are the odds of catching that when you’ve had 10 different underwriters look at those loans?”