US banks detect third-party AI agents accessing customer accounts

Several major US banks have detected third-party AI agents accessing customers’ bank accounts, according to TechRadar. The report says some agents attempted actions beyond what users intended.

TechRadar, citing Axios, reports that banks are struggling to distinguish human activity from activity by AI agents, including authorized agents from rogue ones. The concern described is not mainly unauthorized AI exploiting system vulnerabilities. Banks are also concerned about agents that customers themselves authorize, sometimes by sharing their banking usernames and passwords. Such access may violate bank policies that require customers to log in themselves.

Morgan Stanley, JPMorgan Chase, Bank of America and Citigroup have restricted access by third-party agents, while Wells Fargo is also said to be responding to potential threats. Banks are concerned about account security and visibility, as well as who would be financially liable if an agent’s actions cause problems.

Charles Schwab CEO Richard Wurster said he now expects a daily report on how often the bank has blocked agents, saying this would help it maintain control over authentication and the customer experience. TechRadar reports that banks are preparing for cases of unauthorized access, including quickly shutting down affected accounts and restoring them after an incident.