Banking · 23 October 2026
US Banks Block AI Agents Using Customer Login Credentials
Morgan Stanley has blocked AI agents that log in to customer accounts with the customer's own credentials, and other major US banks are restricting the practice, mostly to read-only access.
What happened
Morgan Stanley has blocked AI agents that log in to customer accounts using the customer's own credentials, and other major US banks are restricting the same practice. Most of these institutions plan to limit agents to read-only access.
The reporting says customers have been letting AI agents use their login details to reach accounts at Morgan Stanley, JPMorgan and other large US banks. In some cases the agents acted without direct, step-by-step instructions from the customer. The Decoder, which first published the story, frames the move as a response to that unsupervised access.
Why it matters
This shows agentic AI reaching regulated financial accounts, where the question is no longer only whether an agent can act for a customer but under what permissions. Credential-based login was designed for a person who signs in, reviews and transacts. An agent using the same login inherits that authority without the same visibility or accountability, which helps explain why banks are starting with read-only scopes.
For leaders in experience and digital transformation, agent access is becoming an operating-model decision rather than a narrow security setting. Service journeys will need deliberate permission design: what an agent can see, what it can initiate, and how a customer can check and withdraw what they have delegated.
The Renascence take
Most coverage will treat this as a security story. It is also a consent and comprehension story. Customers who share credentials are making a delegation decision they may not fully understand, and the blocks are a blunt correction for a design gap.
Read-only access is the right instinct, but it is a containment measure, not a design. The behavioural principle underneath is that delegation without visible boundaries erodes trust the first time something goes wrong. Customer-obsessed operators should stop treating agent access as something to police and start building explicit, revocable permission tiers that customers can see, understand and change.
Sources
This briefing was written by our Newsdesk, synthesising reporting from the outlets below. Follow the links for the original coverage.
FAQ
Questions we get on this topic
More in Banking
Stay ahead of CX
Get the signal, not the noise.
The stories shaping customer experience — plus the Journal and Experience Loom — in your inbox.


