Mortgage fraud risk rises again as new category takes top spot

AI, patterns and new products

Document alteration with AI is something that keeps coming up with clients, Seguin said. He recalled one catch from the subprime crisis.

“A lot of the human telltale signs of document identification are gone,” Seguin said. “I remember looking at a bank statement, and it said ‘Withdrawal maid in store,’ and it was spelled M-A-I-D, and it was one of the big banks. And I knew, ‘Okay, that is not how they’re going to spell the word made.’ And AI’s pretty much taken that away, and those kinds of little signals that underwriters are used to looking for.”

Seguin has been a proponent of stepping back to look for patterns in the data, such as a loan officer whose loans all do the same thing.

“Every pay stub in this whole batch, they all use the same format. And there are 20 different employers here. What are the chances of that?” Seguin said. “It’s going to be harder to identify on a one-off loan basis.”

Non-agency loan types bring their own exposure, from DSCR loans, where borrowers are often LLCs and conflicts of interest can go undisclosed, to bank statement loans, which open the door to bank statement fraud.

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *