United Airlines has announced it will restrict the use of artificial intelligence agents for booking flights and managing reservations, citing significant cybersecurity risks and potential conflicts with federal aviation regulations.
The airline stated that allowing personal or third-party AI assistants to access customer accounts can lead to the exposure of sensitive information. The carrier warned that actions performed by these automated tools could result in unintended purchases, unauthorized account modifications, or the misuse of MileagePlus miles.
“Sharing account credentials with personal, third-party, AI assistants can expose sensitive information,” United said in a statement. “Actions taken by these tools may result in unintended purchases, account changes, or use of MileagePlus miles.”
Additionally, the airline noted that federal rules require a human to complete the flight check-in process to acknowledge that checked bags do not contain prohibited substances. The use of automated agents could undermine these mandatory safety protocols.
As part of its new measures, United is updating its booking, ticketing, and distribution policies to prohibit access by unauthorized third-party personal AI agents. These restrictions apply both to the airline’s direct digital platforms and to indirect channels, including travel agencies and aggregators.
United joins a growing number of major U.S. corporations seeking to prevent customers from using autonomous assistants to access their systems. This move follows a recent warning from Bank of America, which urged customers not to share banking passwords with AI tools last week. The banking sector and airlines are increasingly grappling with the challenge of distinguishing between genuine human customers and AI systems operating on their behalf.
Bank of America did the same thing last week. It is becoming a clear trend across industries now.
I wonder if this applies to voice assistants too. I always worry Siri might book something weird.
Good for United. Security breaches via third-party apps are a nightmare for customer data privacy.
Is this just a delay tactic? In five years, regulators will have to allow automated agents anyway.
The federal check-in rule makes sense. Humans are responsible for knowing what is in their checked luggage, not robots.
Finally, common sense! I do not want my chatbot accidentally burning my frequent flyer miles on a wrong flight.