Citigroup and Bank of America have restricted artificial intelligence agents from accessing customer accounts as US banks seek to protect deposits from automated tools that could move billions of dollars into higher-yielding savings products.

Citi said it would not allow “third-party AI agents or assistants to log into a customer’s account or transact directly within our digital banking platforms”, citing the need to protect customer accounts and information. The bank said customers could still authorise limited, read-only access to account information through certain open-banking providers.

Bank of America chief executive Brian Moynihan confirmed the lender was restricting access while it developed protocols to establish which party would be liable for payments authorised by artificial intelligence. “We strongly believe there should be a human in the loop to authorise money moving out of your account,” Moynihan told CNBC on Wednesday.

The restrictions follow the September launch of Meta’s Muse AI personal assistant, which intensified concerns that automated tools could transfer customers’ idle cash into accounts offering better returns. Technology companies are increasingly developing AI assistants to help consumers manage their finances and make transactions.

JPMorgan, Bank of America and Wells Fargo collectively hold $1.6 trillion in non-interest-bearing deposits, representing 16 per cent of their liabilities. Paying three per cent interest on those deposits would cost the three banks $47 billion annually, almost half their combined yearly earnings, according to the figures cited in Slok’s report.

According to the Financial Times, Wall Street executives have contacted the Federal Reserve over the growing use of AI agents for financial transactions, according to a person familiar with the discussions. The central bank declined to comment, while regulators have privately welcomed banks’ decisions to block access as questions remain over how customers authorise autonomous transactions and who bears responsibility when something goes wrong.

JPMorgan Chase chief executive Jamie Dimon said this week that the bank would restrict outside AI agents because of the risks posed by potentially thousands of rogue systems. The bank is developing tools to protect customers while keeping unauthorised agents out of its systems.

---