The Ban and Its Rationale
On 11 November 2026, United Airlines announced it would prevent artificial intelligence agents from booking flights on its platform. The decision targets tools such as Meta's Muse and OpenAI's "dots", which can autonomously complete purchases rather than simply research options.
The carrier's concern centres on the credentials travellers must hand over for an AI assistant to act on their behalf. United warned that sharing account login details with personal, third-party AI tools can expose sensitive data. The airline also flagged the risk of unintended purchases, account modifications, or the unauthorised use of MileagePlus loyalty miles.
The move reflects a broader tension in the travel industry. Airlines and hotels are trying to balance convenience for customers against control over their own sales channels, all while navigating regulations that were written before autonomous software entered the picture.
A Compliance Problem
United pointed to a specific legal obstacle: the requirement that passengers personally confirm their luggage does not contain prohibited items or substances when checking in for a flight. An AI agent cannot make that attestation on someone else's behalf, the airline said.
That gap between existing aviation rules and the capabilities of new software has prompted United to restrict access not only on its own digital platforms but also indirectly through travel agencies and aggregators. The aim is to block any unauthorised third-party AI assistant from completing transactions.
The issue is not hypothetical. Earlier in 2026, some travellers using American Airlines' own AI agent, AURA, discovered their bookings had been altered without their knowledge. The incident underscored how automation can produce outcomes passengers did not intend or approve.
The Wider Industry Response
Other US carriers are taking different approaches for now. Delta, American, and JetBlue allow customers to book flights through agentic AI platforms, but they place the responsibility for any errors or unintended changes squarely on the traveller. That stance shifts the legal and financial risk away from the airline, though it leaves passengers to manage the consequences if something goes wrong.
The divergence in policy suggests the industry has not yet settled on a common framework. Some carriers appear willing to accommodate the technology while protecting themselves contractually. United, by contrast, has chosen to cut off access altogether until the security and compliance questions are resolved.
What Travellers Are Willing to Hand Over
Consumer appetite for AI in travel is growing, but it is also selective. A survey of more than 3,000 people in Europe and the United States, conducted in summer 2026 by hospitality technology firm Mews, found that over half of respondents were comfortable using AI for translations, recommendations, or summarising reviews. Fewer than half, however, would let an AI plan or book an entire itinerary.
The pattern suggests travellers trust the technology for discrete, low-stakes tasks but remain cautious about handing over decisions that involve money or complex logistics. That hesitation may explain why airlines are proceeding carefully, even as the software becomes more capable.
The Intermediary Problem
United's decision also reflects a strategic calculation. Aviation analyst Henry Harteveldt of Atmosphere Research Group told CBS News that the carrier may be reluctant to introduce another powerful intermediary into the booking process. Airlines already contend with Google search, price-comparison sites, and online travel agencies, all of which sit between the carrier and the customer.
Adding AI agents to that list could further erode the airline's ability to control the customer relationship, influence purchasing decisions, or protect revenue from ancillary services such as seat selection and baggage fees.
Ryanair has fought similar battles with third-party booking platforms, arguing that they mislead customers and withhold booking details, making it harder for the airline to communicate directly with passengers. The European budget carrier has taken legal action to shut down what it calls screen-scraping sites that do not offer the best prices.
The Upsell Dilemma
There is another dimension to the problem, one that may ultimately matter more to airlines than security or compliance. AI shopping assistants are task-oriented and efficient, which makes them better than human travellers at resisting upsells. An algorithm instructed to find the cheapest flight will ignore prompts to upgrade to premium economy or add travel insurance.
For airlines that rely on ancillary revenue to remain profitable, that efficiency is a threat. If a significant share of bookings shifts to AI agents, carriers could see a drop in the add-ons that now account for a large portion of their income.
United's ban may be framed in terms of security and legal compliance, but it also preserves the airline's ability to market directly to passengers during the booking process. An AI intermediary would strip out that opportunity.
What Happens Next
The question now is whether other airlines will follow United's lead or continue to allow AI bookings while shifting liability to the customer. Much will depend on how regulators respond. If aviation authorities clarify the rules around automated check-ins and credential sharing, carriers may feel more confident accommodating the technology.
Travellers, meanwhile, are likely to keep experimenting with AI tools, especially as the software becomes more sophisticated. The tension between convenience and control is unlikely to resolve quickly, and the industry's patchwork of policies reflects that uncertainty.
United's decision is a signal that at least one major carrier believes the risks outweigh the benefits for now. Whether that calculation holds as the technology matures remains to be seen.








