United Airlines, a major U.S. airline, announced on November 11 (local time) that it would prohibit the use of its services by unauthorized third-party personal AI agents or AI assistants. In addition to directly restricting their use on its website and app, it will also restrict use through travel agencies and booking sites (aggregators). It will also revise its policies on reservations, ticket issuance, and sales. The company did not specify which AI agents would be covered or when the restrictions would take effect.
As for the reason for the ban, the company explained that confidential information could leak if account credentials were shared with an AI assistant. The use of AI tools could also result in unintended purchases, account changes, or the use of miles in its MileagePlus mileage program. It further noted that automated operations by AI could violate regulations of the U.S. Federal Aviation Administration (FAA) or the U.S. Department of Transportation (DOT). As an example, it stated that during check-in, a human must confirm that checked baggage contains no prohibited items such as explosives.
The company plans to update its “Digital Trust Center,” which consolidates security information. The updated center will explain the impact of the new restrictions on customers, how to protect their accounts, and where to seek assistance if an AI agent performs unintended actions. In its statement, the company said, “Business leaders in the banking and retail industries are also taking similar measures regarding AI agents,” and called on other companies to consider adopting similar measures.
In the United States, a growing number of companies have sought to shut out external AI agents since Meta released the personal AI agent “Muse” in September. In retail, Amazon began blocking shopping through Muse on September 20. It cited Meta’s failure to provide advance notice, Muse’s practice of accessing Amazon without identifying itself as an AI agent, and the belief that it collects and stores customers’ login information.
In the financial industry, from October 13 through October 20, JPMorgan Chase, Charles Schwab, Citigroup, and Wells Fargo either blocked external AI agents that log in using customers’ own credentials or contractually shifted the risk to customers. JPMorgan CEO Jamie Dimon said during an earnings call that the bank would not allow external agents such as Muse and Instinct to operate within customers’ accounts, citing concerns about “runaway agents” (U.S. CNBC). Bank of America also emailed customers on October 27, urging them not to share their credentials with AI assistants and warning that customers would be responsible for any transactions resulting from such sharing (U.S. TechCrunch).
