America's Biggest Banks Are Blocking AI Agents From Logging Into Customer Accounts. 1Password Now Signs an Agent In Without Showing It the Password.

America's biggest banks have spent the last ten days shutting the door on personal AI agents, software that logs into a website and carries out a task on your behalf. Citi, JPMorgan Chase, Bank of America, Charles Schwab and Morgan Stanley have all said they now block or restrict third-party agents from logging into customer accounts.

Citi's statement, given to Retail Banker International on 19 October, is the plainest: "To help safeguard customer accounts and information, we do not permit third-party AI agents or assistants" to "log into a customer's account or transact directly within our digital banking platforms." Customers can still authorise limited, read-only access through Citi-approved open banking providers.

JPMorgan's Jamie Dimon told analysts on the bank's 13 October results call: "you don't want rogue agents and there could be thousands of them." A Chase spokesperson's advice to customers was simpler still: do not share your username and password with anyone.

Bank of America's Brian Moynihan told CNBC: "We strongly believe there should be a human in the loop to authorise money moving out of your account." The bank is working out who is liable when an AI authorises a payment.

Schwab's chief executive Rick Wurster said the firm is blocking unauthorised agents that try to move clients' money, including Meta's Muse, and that attempts by the big agents run "in the tens and twenties" a week. Schwab is building a controlled connection that would let an agent read balances and do research but not move money.

On 22 October Axios reported that some customers had handed agents their login details, that some agents had contacted bank call centres, and that banks are drawing up contingency plans in case an AI incident locks customers out. Reuters said it could not independently verify the specific incidents.

The money behind the worry: the Financial Times reported that JPMorgan, Bank of America and Wells Fargo hold $1.6 trillion in deposits that pay no interest. If agents moved that cash to accounts paying 3%, it would cost the three banks about $47 billion a year. Wall Street executives have raised the issue with the Federal Reserve.

1Password announced 1Password for Agents on 22 October. It lets a personal AI agent sign in to a website using a password from your vault without the agent seeing the password or the one-time code.

The flow: the agent reaches a sign-in page and asks for a credential. You get a prompt on your own device to pick the right login, you approve it in 1Password, and 1Password fills the form itself. You can disconnect any agent at any time.

Meta's Muse and SpaceXSI's Grok Bot are rolling out over the coming days. Anthropic's Claude Duet, OpenAI's ChatGPT Work and dots, and Town are listed as coming soon, along with a dozen smaller agent tools.

It is free on 1Password Individual and Family accounts hosted on 1Password.com. The post does not mention Business or Teams plans.

Google announced on 23 October that Gemini Notebook, its tool for keeping research notes and source documents in one place, is now a connected app inside the Gemini app.

Type @Gemini Notebook in a chat to create a notebook, add or update sources, save notes, or ask a question that is answered from the notebook's contents. Gemini can also decide to use the notebook without being asked.

The same commands can delete notebooks and delete sources.

It is available now to all Workspace customers, Workspace Individual subscribers and personal Google accounts, on Rapid and Scheduled Release domains. Admins control it through the Gemini and Gemini Notebook access settings, and it only appears for users who have both switched on. Users can turn it off under Settings, then Connected Apps.