United Airlines wants you to check in, not your chatbot

United Airlines has recently implemented restrictions on unauthorized third-party AI agents, warning that these bots could potentially make unauthorized purchases, expose customer data, or circumvent necessary human verification procedures. This move affects both the airline's own digital platforms and indirect access through travel agencies and aggregators.

The restrictions are a part of United Airlines' revised booking, ticketing, and distribution policies. According to the airline, automated actions could potentially violate aviation regulations, specifically the requirement for a human acknowledgment when checking in to ensure bags do not contain any prohibited items. The airline is also concerned about the risks associated with sharing account credentials with personal AI assistants, which can expose sensitive information and result in unintended actions, such as purchases, account changes, or misuse of MileagePlus miles.

United Airlines prefers customers to use its own app rather than relying on third-party assistants. The use of AI agents has become increasingly popular, with influencers often promoting their benefits. However, the issue lies in the proper management and control of these assistants. Earlier this year, Professor Hannah Fry, a British mathematician, conducted an experiment where she granted an AI agent, named Cassandra, excessive control over a bank account.

The results were alarming, with sensitive information being shared and unauthorized transactions taking place. Other major US banks have also followed suit, blocking third-party AI agents from operating within customer accounts. This trend is not limited to the United States, as several UK banks have also restricted the use of these AI agents.

United Airlines has emphasized that other corporate leaders in the banking and retail industries have taken similar actions, urging others to take measures that safeguard customer data and enhance safety. The airline has not yet disclosed its specific implementation plans to block unauthorized AI agents.