Major U.S. banks, including Morgan Stanley, JPMorgan Chase, and Bank of America, are drafting contingency plans to address potential AI-related operational disruptions. The initiatives focus on protecting customer accounts, restoring access, and maintaining operations in the event of an AI-driven incident. Institutions are restricting third-party AI agent access and revising internal rules amid concerns that autonomous agents could execute unauthorized transactions using customer login credentials. Banks note difficulty distinguishing between user-approved actions and unsanctioned activity initiated by AI tools. However, regulators and bank officials confirmed no evidence currently exists of runaway agents or hacker attacks compromising customer funds.