Bank of America, JPMorgan Chase, Citigroup and Wells Fargo are restricting third-party AI agents’ access to customer accounts. The move follows incidents of unauthorized or autonomous AI activity that made it difficult for banks to distinguish people from bots. [2, 3, 10] [3, 6, 7]
Banks cite the risk of rogue agents making unauthorized transactions and the unresolved question of who is liable for AI-authorized payments. [1, 4, 7] They also worry that agents could automatically move customer funds from low-interest deposit accounts to higher-yielding alternatives, threatening a key source of bank funding. [1, 9]
Many financial institutions are limiting AI tools to read-only access. Agents can view financial information for services such as budgeting apps, but cannot move funds or execute transactions. [2, 3, 4]