JPMorgan Chase CEO Jamie Dimon says the rapid expansion of autonomous AI tools will make it difficult to prevent them from accessing financial systems, forcing banks to focus on managing the risks.

Dimon warned that AI agents could soon take actions inside customers’ bank accounts that users never approved, raising concerns about financial security, accountability, and who would be responsible if something went wrong.

“You can't stop it. You just have to, like, manage the hell out of it,” Dimon said while speaking at the Economic Club of Chicago earlier this month, referring to the rapid development of AI. He warned that people who give autonomous agents access to their accounts may assume the systems will act in their best interests, even though that outcome is not guaranteed.

The warning comes as major American banks report detecting third-party AI agents operating inside customer accounts. Some customers have provided these tools with their banking usernames and passwords, allowing them to access financial information and attempt actions on the customers’ behalf. Banks have raised concerns that some agents may act beyond users’ instructions or without specific approval for individual actions.

JPMorgan has begun blocking outside agents from accessing customer accounts through banking credentials, while exploring ways to support more secure, authorized data sharing. Dimon has said the bank is not opposed to AI agents themselves, but unrestricted access from outside systems poses risks to both customers and the institution. He has also warned that customers could face greater responsibility when they hand control of financial information or transactions to third-party tools.

Other financial institutions are taking similar precautions. Morgan Stanley has blocked third-party agents from its platforms, and Charles Schwab has reported blocking unauthorized systems attempting to access customer accounts. Banks are particularly concerned about distinguishing legitimate user-directed activity from actions that have not been approved, especially when automated agents use customers’ own credentials.

Dimon has also warned that increasingly capable AI systems could create broader cybersecurity threats, including automated attacks that operate continuously and at a speed difficult for human teams to match. Rather than expecting the technology to disappear, he has urged businesses and regulators to prepare safeguards as its capabilities expand.

For banks, the challenge is to allow customers to benefit from automated financial tools without surrendering control over sensitive accounts. Until reliable safeguards are in place, financial institutions are likely to continue restricting outside agents that attempt to access accounts or initiate transactions without an approved, secure connection.