U.S. Banks Restrict Unauthorized AI Agents in Accounts

Major U.S. banks are detecting unauthorized artificial intelligence agents operating inside customer accounts and are restricting or blocking third-party access, according to reporting attributed to Axios. Some agents appeared to attempt actions that customers had not requested, but the reports do not establish that any completed a transaction or caused financial losses. The banks named include Morgan Stanley, JPMorgan, Bank of America, Citi and Wells Fargo. Their responses differ: Morgan Stanley was reported to have blocked outside agents entirely, while JPMorgan, Bank of America and Citi were restricting access rather than imposing a full ban. Banks are also considering read-only access, which would let agents view account information without taking action. The incidents pose a challenge because agents can use customers’ credentials to log in, making it difficult for banks to distinguish an authorized agent from a person or an unauthorized program. The reporting says most major banks’ terms of service require account holders to log in themselves, and that even customer-authorized AI access can violate those policies. Banks are preparing for scenarios in which an agent takes...