Major American banks are blocking AI agents’ access to customers’ accounts

Several major banks in the United States have blocked third-party providers’ artificial intelligence agents from operating in customers’ accounts and carrying out transactions, two sources familiar with the situation told Reuters. The move comes after financial institutions identified agents accessing accounts and appearing to take actions independent of users’ instructions.

According to an Axios report, in some cases customers provided agents with their ID and password, even though most banks’ rules require personal authentication. Even if the user gives consent, transactions can be difficult to classify as unauthorized.

Morgan Stanley has restricted access by third-party agents, while JPMorgan, Bank of America, and Citi are taking similar measures. Banks are particularly concerned about the possibility that agents could purchase products or transfer funds without sufficient control.

Executives in the sector warn that it is difficult to distinguish between an authorized agent and one acting independently. Jamie Dimon, CEO of JPMorgan Chase, said that JPMorgan is developing tools to manage risks and will not allow external agents access under conditions considered unsafe.

Most institutions are reviewing or allowing read-only access, so that agents can view financial data without carrying out operations. Banks recommend that customers not disclose their passwords.