Bank of America Warns Customers Against Sharing Banking Credentials With AI Agents

Bank of America has started issuing email warnings to customers, explicitly advising them not to share online banking credentials and usernames with third-party AI assistants and agents. The bank stated that any financial losses or liabilities resulting from sharing these credentials will be borne entirely by the customer. This warning highlights a critical liability and security gap as autonomous AI agents increasingly gain cross-application task execution capabilities. Major financial institutions like Morgan Stanley, JPMorgan Chase, and Citi are tightening permissions or blocking unauthorized agents to mitigate fraud and compliance risks. Banks face a major challenge in determining whether a third-party AI agent is operating strictly within user instructions or executing unauthorized actions. Because traditional fraud detection systems cannot easily differentiate these scenarios, banks are pushing the financial risk back onto consumers.