For smaller banks, where resources can be limited, adopting AI comes with its own set of opportunities and challenges. So how are these institutions using AI — and what does it mean for the future of community banking?
Like so many other industries, the banking sector is figuring out how to navigate AI. And the stakes are high. The tech has raised red flags over potential bias, particularly when it comes to lending practices. But it could also make some banking operations more efficient, which would be especially valuable for smaller banks — which tend to operate with fewer resources than their bigger competitors but also offer more personal service, a connection to the local community, and typically lower fees.
The catch? Those banks can’t necessarily afford to pour resources into AI.
A few days ago, Marketplace’s Justin Ho caught up with Dominik Mjartan, CEO of American Pride Bank, a small lender based out of Macon, Georgia. He was on the road talking to AI software.
“As I was driving, I was putting weekly executive scorecards for our executive team meeting this afternoon,” he said.
Those scorecards cover how many loans the bank made last week, how much interest it pulled in, and how much money customers deposited — by branch and by region.
“Which normally would have taken our CFO, or someone, you’re talking a few hours of work, to do a decent job,” Mjartan said. Instead, it took minutes.
He said the bank has been rolling out AI software for annual performance reviews, budgeting, and market research.
Meanwhile, at First Southwest Bank in Colorado, AI is reviewing loan documents and preparing staff for meetings with borrowers.
“So if a customer comes in with a loan application, we want to be able to operate 10 times or more the speed that we did before, and really focus on aspects of the relationship instead of the administrative piece,” said Kristy Esquibel, the bank’s Chief Credit Officer.
But Esquibel said her bank is not using AI to make any decisions about loans. She’s concerned about how the software analyzes data. For instance, what if AI denies an application from a worthy borrower who happens to live in a low-income zip code?
“I think any AI tool that we’d put in place, you’re going to have to ensure that it’s not creating its own bias, or consistent practices that you don’t intend to put in place,” said Esquibel.
She also has concerns about AI’s accuracy and privacy.
“You’re ultimately responsible for what your software and what your vendors are doing, and so I think regulators are going to expect banks to really understand their models to ensure consumer protection,” she said.
Kathryn Judge, a law professor at Columbia, said regulators could take some of those responsibilities away from banks by doing more oversight over third-party software.
“So that way the bank regulators have the ability to say, ‘No, we understand the nature of the services that this third party is providing, and we’re able to monitor those risks directly.’”
In the meantime, banks need to develop their own guidelines for using AI, according to Surjit Chana, with Beneficial State Bank in California. “They have to ground their AI deployments with a set of values and morals.”
Chana said that means factoring in the impact AI could have on labor markets, privacy, and access to credit.
Ensuring AI doesn’t discriminate, for instance, will also help the bank make more loans, he said. “Because you’re actually expanding the market by tapping into and providing your services to a broader set of communities.”
In other words, Chana said that using AI responsibly could help banks make more money.
Apple
Spotify
Amazon
RSS