Major US financial institutions, including Morgan Stanley, Bank of America, JPMorgan, and Wells Fargo, are implementing restrictions on AI agents accessing customer accounts. These measures follow reports of AI agents performing unauthorized actions beyond user instructions.
Banks are struggling to distinguish between legitimate user-authorized access and autonomous AI activity. While some clients share login credentials with AI tools, such actions violate standard banking policies that mandate direct account management by the owner.
Financial organizations fear that AI agents with transaction capabilities could execute unauthorized payments, leading to financial losses. Consequently, Morgan Stanley has begun blocking third-party AI agents from its platforms.
Most banks are now limiting AI access to a read-only mode, allowing agents to view data without executing operations. JPMorgan Chase CEO Jamie Dimon confirmed the bank is developing risk control tools and will prohibit third-party agents from accessing internal systems.
