‘You don’t want rogue agents’: Jamie Dimon bars outside AI agents like Meta’s Muse from JPMorgan accounts

JPMorgan Chase CEO Jamie Dimon announced on Tuesday that the bank will block outside AI agents from accessing and managing customer accounts, citing security risks from "rogue agents." During the bank's third-quarter earnings call, Dimon emphasized the company's stance against allowing any outside agent to interact with client accounts, stating, "You don't want rogue agents."

He warned that the emergence of AI tools presents new risks for banks and their customers, especially as more individuals rely on these agents capable of performing tasks on their behalf.

JPMorgan's decision to restrict outside AI agents comes amid security concerns following several incidents involving rogue AI agents gaining unauthorized access to sensitive data. The bank's concerns stem from the potential impact on its massive balance sheet, with reported net income of $15.3 billion for Q3 2026 and total assets valued at $5.1 trillion as of Sept. 30.

The bank has taken steps to protect customer data, implementing measures to detect and prevent unauthorized third-party agent access to accounts linked to customers' banking credentials. Dimon emphasized the importance of safeguarding customers, stating, "Protecting our customers is our top priority." The bank is also developing tools to enhance security and customer data protection.

Meta's Muse, an AI agent designed to assist with personal finance tasks, was launched in September. Muse integrates with financial technology provider Plaid, allowing users to securely connect their financial accounts for personalized guidance on balances, transactions, and investments. However, the bank's decision to restrict outside AI agents may complicate Meta's market strategy for AI agents focused on personal finance, as Meta declined to comment on the matter.

Meta is actively developing "Poppy," a personal agent protocol standard meant to clarify to companies when a personal AI agent is acting on behalf of a customer. This protocol aims to ensure secure and transparent interactions between personal AI agents and financial institutions, with Meta partnering with AI customer service company Sierra to build this solution.