U.S. banks move to restrict AI agent access over account security risks

Traders Union
All News
Financial Institutions
US banks restrict AI agents

Highlights

Major U.S. banks including Morgan Stanley, JPMorgan, Bank of America, Citi, and Wells Fargo are restricting third-party AI agent access due to account security concerns.
Banks are shifting toward read-only access models for AI agents, allowing data viewing but blocking transactions, to balance customer adoption and risk control.
Financial institutions are increasing investments in defensive technology and contingency planning to address persistent difficulties distinguishing between authorized users, bots, and potentially rogue AI agents.

Banks tighten access rules as agent activity spreads

As first reported by Axios, major U.S. banks are detecting third-party AI agents inside customer accounts, including cases where the tools appear to act beyond direct user instructions. The reported activity includes agents entering accounts with customer-provided login credentials, a practice that conflicts with many banks' terms requiring customers to log in themselves.

Banks including Morgan Stanley, Bank of America, JPMorgan, Wells Fargo and Citi are now moving to contain the risks. Executives cited concerns that an agent with account access could make purchases, move funds or take other actions that affect a customer's finances.

Some banks are also tracking agent activity through call centers, where automated tools programmed to complete tasks for users are contacting staff. Executives described the scenario as alarming because banks often cannot clearly determine whether the activity is user-directed, authorized or potentially rogue.

Morgan Stanley has begun blocking third-party agents from its platforms, according to people familiar with the decision. JPMorgan, Bank of America and Citi are also restricting agent access, while a Chase spokesperson said customers should not share usernames and passwords with anyone; a Morgan Stanley spokesperson said the bank has not observed rogue AI agents accessing its platforms.

Security burden grows across the banking sector

Most banks have authorized, or are moving toward, read-only access models that would let AI agents view financial information without taking action. That approach mirrors existing arrangements for third-party finance apps such as Quicken, but bank and industry executives say broader AI agent use requires a new balance between convenience and control.

For now, banks see the risks of permissive access as too high, even as they expect customers to keep adopting AI assistants. One executive said lenders struggle not only to distinguish humans from bots, but also to tell whether a particular agent has been approved by the customer or has gained improper access.

The challenge is emerging despite the sector's large technology budgets and experience operating critical infrastructure. Sources said banks are investing in defenses and in contingency planning for worst-case outcomes, including preparations to secure or restore customer access quickly if agent-related incidents escalate.

In our earlier report, we covered the Bank of England’s push for banks to develop “bare metal” recovery options that can keep critical services running if primary systems are compromised. The article highlighted regulators’ concern that advances in AI are amplifying the scale and complexity of cyber attacks, making end-to-end resilience and recovery planning as important as perimeter defenses.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.