Dimon Says AI Agents Paying Bills Need Bank Codes and 'May Not' Act Properly, Calls AI a Tsunami

JPMorgan Chase CEO Jamie Dimon spelled out his objection to outside AI agents in more detail on Wednesday, a day after the bank's third-quarter earnings call where he said JPMorgan would not let them in.

Speaking at The Economic Club of Chicago, Dimon said an agent that makes payments for a customer needs that customer's bank code. "That means they are in your bank and you're assuming they're going to act properly on your behalf and they may not," he said.

He added that agents could do things that are "treacherous" for the bank. He pointed to consumer assistants such as Meta's Muse, which launched in September to handle emails, travel bookings and calendars.

Liability moves to the customer

On the earnings call, Dimon framed the problem as one of responsibility. "There's a liability transfer," he said. "They should understand the liability transfer. If data is being misused, they should understand it's being misused."

He warned there "could be thousands" of rogue agents. "We're building tools to do that, too," he said of customer protection. "In the meantime, we're not going to allow just any outside agent to come in inside the company because it's too risky for the clients and for JPMorgan."

Dimon also told regulators to get ready. "With all these bots and agents and stuff like that, could someone get hurt and cause a problem? Maybe. If I was a regulator, I'd try to be prepared for that," he said. He said they should consider how agents could affect deposits and liquidity risk management.

JPMorgan's own agent is still a pilot

JPMorgan is not sitting out the technology. CFO Jeremy Barnum said on the earnings call that the bank's Smart Cash tool, which moves customer money between accounts at the bank to maximize yield and manage cash flow, is in pilot.

The cost picture is murkier. Last year Dimon said the bank's annual AI-generated savings of $2 billion roughly equaled what it spends on the technology. Asked for updated figures, Barnum put annual AI spending closer to $700 million, while calling that number "a little stale at this point."

He said the spending is hard to isolate. "It's, at this point, not really something that you can ring-fence," Barnum said. It is "pretty well diffused inside the whole institution."

The competitive question

The security argument sits alongside a business one. McKinsey warned in April that retail banks may face disintermediation of their customer relationships if AI agents acting for consumers move money in search of better rates.

Bank of America Securities analyst Ebrahim Poonawala put it this way in a Sept. 23 note: "The strategic question is not whether AI pulls deposits out of checking, but who owns the agent, the wallet and ultimately the customer relationship."

Dimon did not frame his position that way. His stated reason is risk to clients and to JPMorgan.

"Manage the hell out of it"

In Chicago, Dimon said AI is moving far faster than the Industrial Revolution did. "This is doing stuff today that no one would have imagined was possible literally five weeks ago," he said. "It's more like a tsunami. It's rolling through the system, and you can't stop some of it."

His reasoning: open-source models and systems built in China and other countries mean no government can simply halt the technology. "You can't stop it. You just have to, like, manage the hell out of it," he said.

He said AI tools are surfacing hundreds of significant vulnerabilities in open-source code, leaving "industries completely unprotected" against state actors, terrorists and criminals. He said agents could also run cyberattacks around the clock.

Bank of America CEO Brian Moynihan told analysts this week that his bank hired 1,000 external consultants to shore up software vulnerabilities exposed by top-tier models such as Anthropic's Mythos.

Reuters reported that the Federal Reserve, the Office of the Comptroller of the Currency and the FDIC are questioning lenders about advanced AI systems and safeguards. It also reported that the Bank of England and the European Central Bank are pressing banks to improve recovery plans for AI-related disruptions.

The optimistic case

Dimon was upbeat about the payoff. He said AI could drive breakthroughs in medicine and transportation, and predicted it will help treat diseases such as cancer and cut road deaths by as much as 80%. In 30 years, he said, people may work three and a half or four days a week with per-person income of $300,000.

Those are Dimon's predictions, not findings. They come as Google, Amazon, Microsoft and Meta planned more than $700 billion in capital spending this year, much of it on AI infrastructure.

It is unclear when, and on what terms, JPMorgan will let outside agents past the front door. Dimon has said the bank is building protection tools first, and the Smart Cash pilot is the only agent-driven money movement the bank has described so far.