Upstart: Humans are not very good at underwriting loans so AI won’t be either

“…unfortunately, humans have never really been very good at precisely underwriting loans and figuring out the cash flows they’re going to produce for the next 5 years,” said Upstart CEO Paul Gu during the company’s Q4 earnings call in response to an analyst’s question. “That’s something that has always been solved as a big math problem.”

Upstart’s innovative consumer credit models preceded the dawn of modern-day LLMs. It has been one of their defining features. Underwriting on their part is a combination of the best data access and math. Because of that, they do not view AI as a threat because AI is only great at replacing what humans are good at and underwriting is not one of those things.

“I mean the simple answer is just that a lot of the advances in AI are really good for work that humans are naturally good at,” said Gu.

Gu used an example of a HELOC in which human processors have to go through process of securing and perfecting a lien, checking property records, etc. “…like a lot of that stuff is a mess in a human way and traditionally comes with very high operations cost because you have a lot of people that are checking to make sure things are right,” Gu said. “Those are actually the perfect problem to throw sort of LLM-style AI against.”

When it comes to AI benefitting their business, that’s how Upstart is approaching it.

“…It’s really important to just remember that the LLM models coming from Anthropic or OpenAI or any of the others, Gemini, they are really good at solving problems that humans are good at solving and they can do it at scale. They can work 24/7. You can spin up 100 of them in parallel and have them work. But no matter how many humans you have, you don’t want that army of humans underwriting loans for you,” Gu said.