Your AI can plan the trip. Who lets it buy the ticket?

Anyone who has compared fares across a dozen tabs can understand the appeal of an assistant that does the searching. The harder moment arrives when searching becomes spending: which flight was authorized, what baggage conditions were understood, and who can undo the mistake? United's November 11 statement says it is restricting unauthorized third-party AI agents on its own platforms and indirectly through travel agencies and aggregators. It cites customer-data and account risks, unintended purchases and changes, and compliance concerns. Those are the airline's stated reasons, not findings that every third-party assistant has caused harm. This is not a ban on all uses of AI in travel. The supplied AI News article reports a different emphasis from JetBlue president Marty St. George: airlines should respond if customers prefer AI intermediaries. It also makes an important distinction—JetBlue has not announced a policy authorizing third-party agents to complete bookings. His concern about future intermediary charges points to a commercial dispute as well as a security question. Customers should not have to choose between useful assistance and meaningful control. A better design would separate searching from purchasing, require specific consent for consequential changes, and make disputed transactions traceable. Those are proposed safeguards, not features we have verified in these services. Restrictions may protect travelers; they can also entrench the company that controls access. The test is whether an airline offers a workable, secure route for legitimate delegated use rather than merely closing a gate. Equally, an agent provider should not treat a user's general request for a holiday as unlimited authority over their account.