United Airlines Bars AI Agents From Booking Flights in First for US Carriers

United Airlines has become the first US carrier to bar unauthorised third-party AI agents from booking its flights, drawing a hard line between software that helps travellers plan a trip and software that completes the purchase for them.

The Chicago-based airline announced the policy on Wednesday, 11 November, and wrote it into its contract of carriage the same day. From now on, an AI assistant may look up United flights and show fares, but it may not buy the ticket. The traveller has to finish the booking.

What the ban covers

The restriction applies across United’s own website and app, and indirectly through travel agencies and aggregators, so an agent cannot get around the rule by routing the booking through a middleman. United said it is updating its booking, ticketing and distribution policies to “prohibit use by unauthorized, third-party personal AI agents or assistants”.

The revised contract of carriage defines an AI agent as software that books, buys or manages a United product on its own. A new clause, Rule 6.J.6, lists “the use of AI Agents to make, modify or service a Ticket” among prohibited ticketing practices. A separate clause, Rule 5.H, gives the airline the right to cancel reservations made, modified or serviced by such software.

Crucially, a customer’s own permission is not enough. United’s guidance says an assistant cannot transact with the airline merely because its user has granted it permission; the tool must also have United’s authorisation. The airline says it does not currently work with any third-party personal AI agents, but it wants to build “standardized, secure technology integrations” with privacy, authorisation and accountability protections first.

Why United says it is acting

United gave three reasons, all centred on what happens when software holds a traveller’s logins and acts on their behalf. Sharing account credentials with a personal, third-party AI assistant “can expose sensitive information, and actions taken by these tools may result in unintended purchases, account changes, or use of MileagePlus miles”, the airline warned.

The second concern is regulatory. Online check-in requires a passenger to personally acknowledge rules on prohibited items in checked baggage, including explosives, under federal rules (49 CFR 175.25). United said an automated agent completing that step could create a compliance risk with Federal Aviation Administration and Department of Transportation requirements.

The move follows a warning from banks, including Bank of America, which last week urged customers not to share their passwords with AI assistants. It also lands as AI travel tools are moving fast in the opposite direction: OpenAI’s acquisition of AI travel agent startup Odessia became public a day earlier, and Airbnb’s chief executive Brian Chesky discussed the company’s AI plans on its third-quarter earnings call.

What travellers can still do

The ban is on transactions, not information. AI assistants can still search United flights, compare fares and help plan a trip, as long as they do not touch United’s website or app to complete the purchase. Assistive technology for customers with disabilities is unaffected.

Travellers who have already used an agent get a specific reassurance: bookings an AI agent made before 11 November will not be cancelled on that basis, although United recommends checking that recent trips were booked and ticketed correctly. From now on, however, a reservation made, modified or serviced by an unauthorised agent can be cancelled under Rule 5.H.

The commercial question behind the safety case

United’s public case is about security and compliance, but its chief executive, Scott Kirby, was frank about a commercial motive at an industry event in Chicago this week. “To be honest, I don’t want to create another middleman,” he said, according to the Wall Street Journal. If an AI agent books the flight, United loses the chance to sell the traveller add-ons and upgrades directly, and with it the direct relationship with the customer.

That framing matters because the policy is United’s alone for now. Delta Air Lines has not followed, and no other large US carrier has announced a similar ban. Whether rivals copy United — or whether approved, standardised integrations emerge instead — will shape how much of the booking market AI agents can reach. For now, United has decided the checkout page is its own, and an agent that wants in will need the airline’s permission, not just the traveller’s.

Frequently asked questions

Can an AI assistant still book a United flight for me?

No. From 11 November, an AI assistant may search United flights and show fares, but it may not complete the purchase, change a booking or check you in. You have to finish the booking yourself.

What happens to a trip an AI agent already booked?

Bookings made by AI agents before 11 November 2026 will not be cancelled on that basis. United recommends reviewing recent trips to confirm they were booked and ticketed correctly.

Why does United say the ban is needed?

The airline cites three risks: sharing account credentials can expose sensitive information; an agent’s actions can lead to unintended purchases, account changes or use of MileagePlus miles; and automated check-in could conflict with federal rules requiring a passenger to personally acknowledge dangerous-goods information.

Does the ban affect fare searches or trip planning?

No. AI tools can still show United flight options and prices, and travellers can use them to research and plan, as long as the tool does not access United’s website or app to transact.