AI assistants may soon run your financial life. That could be a big problem for banks.
Americans forgo untold sums of money every year on what some might call a laziness tax, for oversights like failing to refinance their mortgage or incurring a late fee after forgetting to pay a bill. Another silent wealth killer: keeping too much cash sitting in checking or savings accounts that pay little or no interest.
Now, artificial intelligence is threatening to put an end to all that.
Apollo Global Management Chief Economist Torsten Slok recently raised the prospect of an AI-induced bank run, saying that people could use bots to sweep their money into accounts that pay higher interest rates.
Slok’s observation was widely shared on X, and brought to the fore other ways AI could save people money—and hit corporate profits.
“Consumer inertia has been a very powerful force in shaping the financial industry,” said John Campbell , a professor of economics at Harvard University.
Examples abound across finance where AI might help people save money, including on credit card payments, overdraft fees and mortgages.
Morgan Stanley Research said it estimates that AI could double the portion of borrowers refinancing or prepaying their mortgages. Homeowners might use AI to monitor rates, for instance, and automatically prepare the necessary documentation for refinancing. Already, large mortgage lenders including United Wholesale Mortgage have AI tools to monitor when clients could save money on a refinance, and call them to pitch it.
Lenders might benefit as well from increased volume, Morgan Stanley Research said.
Campbell noted that lenders might also try to offset lost interest revenue by raising the mortgage rates they offer in the first place.
The prospect of losing deposits sitting in checking and savings is especially troubling for banks, which rely on those funds to make loans that help keep the economy humming, pocketing the extra interest they make.
“Banks may call it an agentic bank run. Customers might call it checking the interest rate,” said one user on X, in response to Slok’s observation.
About $7.12 trillion sits in consumer and business bank checking accounts, according to the Federal Reserve Bank of St. Louis. Those often bear little or no interest.
Many Wall Street analysts don’t think a draining of deposits would happen overnight, if at all.
For one, banks’ larger institutional clients are largely already doing this as a part of so-called treasury management, which involves moving cash into higher-yielding accounts or using it to pay down debt. Consumer accounts may only have a few thousand dollars in balances, and thus relatively less to gain.
There’s also the trust factor. People like having their money somewhere they feel is secure, which is often in the traditional banks they know well. Big banks, for their part, tend to covet customers’ direct-deposits, sticky funds that rarely move or chase rates, while accounting for the chance other deposits are flightier.
“You’re not going to give your money to some bank you’ve never heard of,” said Peter Crane of Crane Data, which researches money-market funds.
Still, investors appear to be taking the prospect of AI disruption seriously. Financial stocks, including consumer banks large and small, slid last week following Meta Platforms ’s release of its Muse virtual assistant.
Meta has promoted Muse as an instant money saver, with the ability to help renegotiate your utility bills or spot subscriptions you’re not using. It can act autonomously on a person’s behalf (though it can’t currently move account balances on its own) and has sponged up users since its debut.
The idea that AI agents could level the playing field in consumer finances relies on the assumption that bots will recommend the best product or money-saving tactic. But they could have blind spots.
There’s also the possibility that recommendations could one day be influenced by sponsorships, said Campbell, the Harvard professor.
“These conflicts of interest don’t go away just because of the technological advance,” he said.
Write to Ben Glickman at ben.glickman@wsj.com
