Effective immediately, United Airlines is rewriting its ticket-sale policies to prevent artificial-intelligence agents from completing reservations on its own website, through external travel agencies, and via flight-search aggregators. The move mirrors steps taken by other large corporations, such as JPMorgan Chase and Amazon, that have blocked AI-driven interactions to safeguard customer data and maintain compliance with existing regulations.
United’s security team argues that handing over sensitive information—payment details, loyalty-program credentials, or even simple login tokens—to an autonomous software agent creates a vector for unintended purchases, unauthorized account changes, or the misuse of MileagePlus miles. The airline also points out that certain mandatory disclosures, like confirming that checked baggage contains no prohibited items, are legally required to be made by a human person under Department of Transportation (DOT) and Federal Aviation Administration (FAA) rules.
Regulatory fallout and the need for human verification
Federal regulations make it clear that only a living individual may affirm that a bag is free of explosives during the check-in process. By allowing an AI system to perform that affirmation, United could unintentionally breach both DOT and FAA statutes. In addition, the airline fears that an automated system could inadvertently expose personally identifiable information, violating privacy standards that govern the airline industry.
“To be honest, I don’t want to create another middleman,” United CEO Scott Kirby said at a recent industry gathering, emphasizing his reluctance to hand control of the purchasing funnel over to a third-party algorithm. Kirby’s comment reflects a broader corporate skepticism about the readiness of agentic commerce for the mass market.
How competitors are handling AI-enabled travel bookings
Not every carrier shares United’s caution. Southwest Airlines, for example, has launched a plug-in that lets passengers ask ChatGPT to locate and book flights, effectively embracing the AI-assisted shopping model. The divergent approaches illustrate a split in the airline sector: some see AI as a competitive advantage, while others view it as a liability.
Beyond airlines, the broader ecosystem shows modest adoption of AI for transaction-level tasks. In its most recent quarterly earnings call, Uber disclosed that during the week surrounding Halloween, its platform processed 320 million trips and deliveries, yet only 500 Uber Eats orders and a mere four rides were placed by AI agents. Those figures suggest that, for now, AI-driven commerce remains a niche activity rather than a disruptive force.
What the policy means for travelers
Passengers planning a United flight will need to complete the entire reservation themselves, whether they start on United’s official site or on a third-party portal that aggregates multiple airlines. The airline advises customers to keep personal data—credit-card numbers, MileagePlus credentials, and email passwords—in their own hands, rather than entrusting it to an autonomous chatbot.
While the restriction may feel like an inconvenience to tech-savvy users, United argues that the trade-off protects both the consumer and the carrier from potential fraud, privacy breaches, and regulatory penalties. As AI assistants continue to evolve, the airline hints that future policies could be revisited, but for now the emphasis is on human-only transactions.
