Banks have a new problem. AI is logging into customer accounts

Banks like Morgan Stanley, Bank of America, JPMorgan, and Wells Fargo are limiting AI agents' access to customer accounts due to security concerns. Some agents logged in with customer consent, but this may violate bank policies requiring customers to log in independently. Issues arise when agents have permissions for transactions. Banks are struggling to determine if account activities were initiated by customers or AI agents. Morgan Stanley has started blocking external AI agents, while others consider read-only access. Regulators are inquiring about banks' controls over autonomous AI systems, including monitoring and auditing capabilities.