United Airlines Bans Autonomous AI Agents From Booking Flights

Travel advisors worried that autonomous AI bots will soon replace them just got some breathing room from one of the world’s largest airlines.

United Airlines on Wednesday said that AI agents, like Meta’s Muse or agents you can build on Claude, will not be allowed to book flights through United. United made the decision because of perceived dangers of exposing sensitive information, potential violations of federal rules, and the likelihood that the agent could make a mistake when booking (other industries, including banking, have taken a similar approach).

The news is good news for advisors who believe AI agents to be a threat to business sometime in the future (the same can be said about flight aggregators and OTAs), but it’s likely not the end of the threat.

Other airlines, like Southwest, already list flights inside of ChatGPT and other LLMs like Gemini and Claude. The difference with that is that actual humans are booking the flights through the platforms instead of AI assistants and agents doing it for them. The user is the one responsible for personal and payment information in those scenarios, instead of an AI agent.

That is currently the hardline for United. As long as a human is making the decision, it doesn’t matter what avenue they take to book their flight, whether that’s going direct to the United website, using an advisor who books via a GDS, or doing so on Google Flights or ChatGPT.

The ban is even written into United’s updated Contract of Carriage. Under the revised policy, travel agencies and booking platforms are prohibited from using third-party AI agents to search for, book, or manage United travel across both direct and indirect channels. Agencies can only deploy their own internal AI tools with explicit permission from the airline.

The question now is what do the other legacy airlines do, including Delta Air Lines and American Airlines, two that have taken a wait-and-see approach so far. The other question is when and if United’s policy changes. Here’s what the airline said about that:

“As personal AI assistants evolve, United will continue to evaluate how these tools interact with our services, with a focus on protecting our customers,” it said in its statement.

It’s still very early for autonomous agents, and, eventually, those agents might be sophisticated enough to dispel those fears about privacy and rules violations.

For now, United’s firewall gives the advisors some breathing room, but it’s clear that the broader technological tide will keep rising. For travel advisors, the ultimate safeguard against agentic AI won’t be airline policy restrictions; it will be the irreplaceable value of human expertise.