Saturday, 14 November 2026
0 agent hacks today▼ 4 vs yesterday (4)

United Airlines moves to block AI agents from customer accounts

The airline says letting third-party AI assistants hold account credentials risks exposing personal data and triggering purchases customers did not intend.

By The Agentic Times ·

United Airlines moves to block AI agents from customer accounts
· Image: therecord.media

United Airlines plans to stop customers using third-party AI agents on its accounts, citing cyber security risks, according to a report published by The Record on 12 November 2026.

In a statement quoted by the outlet, the airline said that "sharing account credentials with personal, third-party, AI assistants can expose sensitive information, and actions taken by these tools may result in unintended purchases." The company listed further concerns alongside those two.

The policy targets a pattern that has become common over the past year: consumers handing login details to an AI assistant so it can search for flights, compare fares, or complete a booking on their behalf. Because most of these tools act by driving a normal web session, they need the same credentials a human would use. From the airline's side, the traffic looks like an ordinary customer logging in.

That is the heart of the problem United appears to be describing. A credential given to an agent is a credential stored somewhere outside the airline's control, often with a provider the airline has no contract with and no visibility into. If that store is breached, the account goes with it. And an agent that can log in can also transact, which is where the airline's warning about "unintended purchases" comes in: an assistant that misreads an instruction can buy a ticket, change a seat or spend loyalty points, and the customer only finds out afterwards.

The Record's report does not say when the block takes effect, how United will detect agent traffic, or whether the restriction applies to all automated access or only to assistants using customer credentials. Nor does the report describe any specific incident that prompted the decision. United's statement, as quoted, is framed around potential exposure rather than a breach that has already happened.

The move puts United among the companies drawing a line between official integrations and consumer agents acting on a user's behalf. The distinction matters for security teams: an approved integration can be scoped, rate-limited and revoked, while a general-purpose assistant holding a password has whatever access the customer has.

For customers who have already shared United credentials with an assistant, the airline's own reasoning points to an obvious step, which is changing the password and reviewing recent account activity. United did not, in the reported statement, offer guidance on that point.

The Agentic Times has not independently verified the policy beyond the reporting cited here.

Sources

  1. therecord.mediahttps://therecord.media/united-airlines-to-block-customers-using-ai-agents