Major U.S. banks are restricting third-party AI agents from logging into customer accounts with user-provided credentials. Axios reports that Morgan Stanley, Bank of America, JPMorgan, Wells Fargo and others are moving to contain security and liability risks after agents appeared to act beyond clear user instructions in some cases.
The policies are not identical. JPMorgan blocks external agents from operating retail accounts or signing in with customer credentials. Citi bars agent logins and direct transactions but allows limited read-only access through approved open banking connectors, according to USA Today.
Builders of AI banking agents should treat credential-based access as a shrinking path. Approved connectors, scoped permissions and human approval before sensitive actions are becoming the safer design.