United Airlines blocks AI agents from booking flights

United Airlines has made it clear that it will not permit personal artificial intelligence assistants to purchase tickets on its airline, pointing to significant privacy and regulatory concerns surrounding automated transactions.

While the company notes it has not yet observed widespread attempts by individuals using AI agents to book travel—matching observations from other companies like DoorDash—executives remain worried that autonomous agents could expose sensitive customer information, make unintended purchases, or potentially bypass federal regulations requiring airlines to share specific information directly with passengers.

This move highlights a broader industry friction as autonomous AI agents evolve to execute tasks on behalf of users. While tech companies like Meta with Muse and OpenAI with Dots are pushing for deeper personal assistant integration, commercial enterprises face complex liabilities regarding account security, fraud prevention, and strict regulatory compliance that current AI frameworks cannot fully guarantee.

A key regulatory hurdle highlighted by the airline involves mandatory legal attestations, such as requiring travelers to personally confirm they do not have explosives in their luggage—a verification step that automated third-party AI agents cannot legally or physically attest to on behalf of a human.

Although these automated agents will still be permitted to search and display United flights to users, they are strictly blocked from completing the final transaction checkout. This restriction directly impacts the utility of newer personal AI tools currently being developed across the tech sector.