United Airlines restricts AI agents’ access over security concerns
United Airlines has restricted AI agents from accessing its digital platforms to book flights, citing security and regulatory concerns.
The carrier announced on November 11, 2026, that it will amend its Contract of Carriage to explicitly bar AI agents from completing any ticket-related transactions. This restriction applies even when a customer is supervising the bot's actions, requiring a human to complete the purchase.
Research Permitted, Purchases Restricted
The new language, found in Rule 6.J.6 of United's Contract of Carriage, maintains a clear distinction between research and transactions. Travelers can still use AI tools to compare fares and look up options, but the actual purchasing must be done by a human.
Security and Compliance Concerns
United outlined three primary concerns: unauthorized purchases, potential misuse of loyalty miles, and regulatory compliance. The airline emphasized the need for human oversight during check-in to avoid violations of FAA and DOT rules on prohibited items in luggage.
The airline wants to avoid creating 'another middleman' between itself and the people who fly with it.
Growing Industry Trend
United is not alone in this approach. Amazon, eBay, and several major banks have also restricted AI agents over security concerns. The policy raises practical enforcement questions, as distinguishing human from AI interactions can be challenging.