Opinion: United Just Banned AI Agents From Booking Your Flights. The Fine Print Goes Further Than the Headline.

CHICAGO - The pitch for personal AI agents has always included travel. Tell the assistant where you want to go, and it finds the flight, picks the seat and pays. United Airlines just said no. On Wednesday the carrier said it is restricting access by unauthorized third-party AI agents, “directly within our digital platforms and indirectly via travel agencies and aggregators.” Agents can still show United flights and prices, USA Today and The Verge reported. You have to finish the purchase yourself.

Some of United’s reasons are good ones, and they deserve a fair hearing. Handing software your MileagePlus login is risky. The airline warns that granting that access “has the potential to expose personal information or result in bookings, account changes, or use of MileagePlus miles without your approval.” It also makes a regulatory point that is hard to wave away: “flight check-in requires human acknowledgment that bags do not contain banned substances like explosives.” An agent cannot honestly make that promise for you.

Image: United Airlines

The problem is how far the contract goes. United’s Contract of Carriage now defines an AI Agent as software that takes autonomous or semi-autonomous action “whether or not Passenger or another individual supervises or approves its actions.” It lists “the use of AI Agents to make, modify or service a Ticket, and/or to purchase any service or product” among prohibited practices. And it adds “reservations made, modified or serviced by AI Agents” to the list of bookings United will cancel without notice. The contract’s general remedies for rule violations, which United can use without notice, run from invalidating tickets to deleting miles, revoking Elite status and ending a customer’s MileagePlus membership.

Read that together and the risk shifts to the traveler. If an agent books a flight you asked for, checked and approved, United has written itself the right to cancel it. That is a strange way to protect customers from “unintended purchases,” the phrase United used in its statement, as quoted by USA Today. The intended purchase is the one at risk.

Image: United Airlines

The motive is not only safety. “To be honest, I don’t want to create another middleman,” United chief executive Scott Kirby said at an event on Tuesday, according to The Wall Street Journal, as reported by The Verge. That is a fair business concern. Airlines have long preferred that customers book direct, and an AI assistant that shops every carrier at once looks like the next intermediary. But a distribution fight should be argued as one, not folded into a security policy that threatens a customer’s miles.

United is not alone. USA Today noted that some banks restricted agent access in October and that Amazon blocked agents from its website in September, and it called United the first airline to announce a ban. Meta’s Muse agent passed 10 million users in October, the paper reported, so this is not a niche fight.

There is a better version of this policy, and United has already described it. The airline says it wants “standardized, secure technology integrations with privacy, security, authorization and accountability protections in place,” and when its own FAQ asks whether United works with any third-party AI agents, the answer is “Not yet.” Give that a date and a public process. Let approved agents book with a required human confirmation step, keep check-in attestations human, and warn customers before canceling anything. United also deserves credit for two details: bookings agents made before Nov. 11 will not be canceled, and assistive technology for customers with disabilities is not affected. Build the rest of the policy with the same care, and travelers will not have to choose between a useful tool and their frequent flyer account.