Major US banks draft AI risk contingency plans to guard against account and operational disruptions

October 24 — According to Axios, major U.S. banks are drafting contingency plans for major AI-related incidents, so as to protect accounts, restore customer access and maintain operations after a potential shock. Institutions including Morgan Stanley, JPMorgan Chase and Bank of America are restricting third-party AI agent access and adjusting related rules.

Banks worry that after customers provide AI agents with their login credentials, the agents may carry out actions without the user's explicit approval, and that banks find it difficult to determine whether such activity was authorized by the user. The report also stressed that banks say they have not found evidence of runaway agents or hacker attacks affecting customer funds and accounts.