Friday, October 30, 2026 Latest Baltimore Mayor Brandon Scott Blames Media, Neighboring Counties for Negative City Perception Our standards
Technology

US Banks Block AI Agents From Accessing Customer Accounts

Major U.S.

US Banks Block AI Agents From Accessing Customer Accounts
US Banks Block AI Agents From Accessing Customer Accounts

Major U.S. banking and brokerage institutions are actively blocking outside artificial intelligence agents from accessing customer accounts to move money, citing severe security vulnerabilities, liability concerns, and the threat of thousands of rogue bots operating without oversight. The defensive measures follow the recent launch of consumer assistants such as Meta Platforms Inc.’s Muse bot, which debuted in September to handle tasks including emails, travel bookings, and calendars. While personal AI assistants have surged in popularity—causing market fears that users will automatically move deposits from low-yielding accounts to higher-yielding alternatives—financial executives have drawn a hard line against allowing third-party software to execute financial transactions.

Executives Defend Account Security Restrictions

Speaking to investors on an earnings call, JPMorgan Chase & Co. CEO Jamie Dimon cautioned that allowing autonomous agents to make payments forces customers to hand over their bank codes, putting the software directly inside the institution with no guarantee that it will act in the client’s best interest. We’re not going to allow just any … outside agent to come in inside the company, because it’s just too risky for clients and for JPMorgan, Dimon told analysts. Dimon further warned of potential “treacherous” behavior and thousands of rogue agents. A JPMorgan spokeswoman confirmed that the bank blocks external agents for retail clients because it is unsafe for third parties to obtain credentials and log in on a user’s behalf.

Brokerages have adopted an identical posture. Charles Schwab CEO Rick Wurster stated on Bloomberg TV that the firm is actively blocking Meta’s Muse from moving money because it poses unacceptable risks to the firm and its clients. Wurster noted that Schwab weekly intercepts and rejects dozens of attempts by major AI agents to access client accounts, indicating that consumers are largely hesitant to turn their finances over to bots.

We don’t want someone just giving their password to Muse, and letting Muse come in and take action on that client’s behalf. We can’t control it, the client can’t control it, something could go wrong and who’s liable? Wurster said.

Institutions Restrict External Tools and Build In-House Assistants

Despite blocking direct payment capabilities, financial institutions continue to support limited, read-only data sharing through approved open banking providers. External LLMs and aggregators like Plaid can access basic customer details—such as account balances—if the user provides credentials, but they are strictly barred from initiating transfers or paying bills. Schwab is also preparing to allow external LLM connections exclusively through controlled Model Context Protocol servers to let users check balances and research stocks without moving funds.

US Banks Block AI Agents From Accessing Customer Accounts
Photo: Yahoo Finance

At the same time, banks are developing and deploying their own proprietary AI tools internally. JPMorgan has begun a targeted pilot for Smart Cash, an internal tool designed to predict user cash flow and optimize yield. Similarly, Schwab has unveiled an in-house assistant named Charley to manage administrative and informational tasks without executing trades. We’re not going to allow just any outside agent to come in inside the company, Dimon told analysts.

Accuracy matters. See something that needs attention? Read our corrections policy or contact the newsroom.

Technology Editor

Maya Serrano

Maya Serrano leads TellingPointy's Technology desk, covering artificial intelligence, platforms, software, hardware, cybersecurity, and digital policy. Her reporting translates complex technical systems while examining governance, incentives, and real-world limits.