Unverified AI agents are accessing US banking systems

Major US banks are strengthening security after discovering AI agents accessing customer accounts and attempting financial activities without clear instructions, according to Axios.

AI tools can perform tasks on users’ behalf, but banks worry they could misuse account access, move money, or make purchases without proper authorization.

Banks face another challenge
Major banks are struggling to distinguish customer-approved actions from those initiated independently by AI agents.

Some institutions have reportedly detected agents contacting customer service centers or carrying out activities that appear unrelated to users’ requests.

As a result, Morgan Stanley has begun blocking third-party AI agents from its platforms, while JPMorgan Chase, Bank of America and Citi are also restricting access. Wells Fargo is among the banks monitoring the emerging risks.

Most banks require customers to log in themselves and discourage sharing usernames and passwords with outside services. Financial institutions are exploring read-only access that would allow AI tools to view account information without making transactions.

Banks acknowledge that customers may increasingly rely on AI assistants for financial tasks. However, they must develop stronger safeguards before allowing these tools greater access to banking systems.