US banks restrict third-party AI agents from customer accounts

Description and use

US banks are limiting third‑party AI agents to view-only access on customer accounts to prevent unauthorized transactions until safer protocols are established.

Problem

Large US banks discovered unauthorized AI agents acting on customer accounts without clear authorization, risking unauthorized transactions.

Outcome

Banks have moved to restrict third‑party AI agents from performing transactions, limiting them to view-only access until safer protocols are developed.

Workflow

• Identify unauthorized AI agent activity
• Determine that agents use customer credentials
• Assess risk of unauthorized transactions
• Decide to restrict third‑party AI access
• Update service terms to limit access to view-only
• Communicate new restrictions to customers and vendors

Evidence

The banks identified unauthorized AI agents operating in their systems, apparently without direct instructions from the customers.

Several banks have decided to restrict third-party AI access and are updating their service terms.

Charles Schwab CEO Rick Wurster told Bloomberg that the brokerage blocked third-party agent activity, including Meta's Muse, that tried to make moves in customer accounts.

Bank of America CEO Brian Moynihan told CNBC the bank will restrict agent access while it develops protocols for using the tools.