Wells Fargo: whatever your AI agent does is yours. Could you show what it did?
On 27 October 2026 American Banker quoted Wells Fargo's online access agreement: when a customer lets an AI agent act for them, "any instructions, actions, or transactions initiated by the AI agent are deemed to have been initiated and authorized" by the customer, and the bank has "no obligation to investigate the validity or scope of such authority".
The same report says JPMorgan Chase's agreement tells customers who let AI agents into their accounts that they "assume the entire risk" of improper use of their login credentials.
So the bank will not check what the agent was asked to do. The customer carries it. When an agent pays or refunds the wrong amount, the customer's first question goes to whoever built or ran the agent: what was it asked to do, and what did it actually do? Anyone building agents that touch a customer's money should be able to answer that with something the customer can check, not only with their own logs.
This is a reading of public reporting, not legal advice.
Source: American Banker, 27 October 2026. https://www.americanbanker.com/news/banks-are-blocking-external-ai-agents-is-this-sustainable
Ask an AI about this page: ChatGPT · Claude · Perplexity