United Airlines has announced that it is restricting unauthorized third-party AI agents from booking flights and managing customer reservations, citing concerns about data security, account access and unintended transactions.
The November 11 decision comes as AI assistants become increasingly capable of making purchases and completing tasks on behalf of consumers.
United warned that sharing account credentials with AI agents could expose sensitive personal information and lead to unauthorized purchases, account modifications or the use of MileagePlus miles.
The airline also raised regulatory concerns, noting that automated check-ins could bypass FAA and DOT requirements for passengers to personally acknowledge that their baggage does not contain prohibited items.
Kevin Surace, CEO, TokenCore
“I think United has two concerns here, and only one is being discussed publicly. The first is commercial. AI agents will become exceptionally good at finding the lowest fares, comparing every possible alternative, and making purchasing decisions entirely in the customer’s interest. Airlines have spent decades optimizing pricing, merchandising, and upselling. An intelligent agent working exclusively for the traveler will disrupt that model considerably.
“The second concern is legitimate liability. What happens when an agent books the wrong flight, spends someone’s MileagePlus miles, or purchases a ticket the customer insists they never authorized? Who pays when the customer disputes the charge? These are real issues, as are security and regulatory compliance. But they are solvable through verified identity, explicit customer authorization for financial transactions, spending limits, and auditable consent.
“Ultimately, I believe United will lose this battle. AI agents will increasingly act on our behalf, just as executive assistants have done for decades. We don’t prohibit executive assistants from booking flights because they might make mistakes. We establish authorization and accountability. The same will happen with AI agents. United’s decision may make sense today, but five years from now, preventing customers from using their own intelligent assistants to book travel will seem as antiquated as requiring everyone to call an airline ticket office.”
ㅤ
John Strand, Owner, Black Hills Information Security
“I think United’s caution makes sense. When you let an AI agent act on a customer’s behalf, you’re giving it access to parts of the site that are only available after login. A malicious agent could look for weaknesses in those areas, where it may be able to do more damage. From the outside, it’s hard to tell whether United is acting out of general caution or has specific concerns about what agents could do once authenticated.
“I also understand the concern about having passengers personally acknowledge that their bags don’t contain prohibited items. That’s a declaration you’d want the passenger to make directly. I expect we’ll see more of this from SaaS vendors. They’ll increasingly treat their authenticated services as walled gardens and decide which agents, if any, they’re willing to let in.”
ㅤIt’s about time that someone did this as AI agents do not have the required controls present. Hopefully others follow this lead.
Related
This entry was posted on November 13, 2026 at 7:30 am and is filed under Commentary with tags AI, United Airlines. You can follow any responses to this entry through the RSS 2.0 feed.
You can leave a response, or trackback from your own site.
United blocks third-party AI agents from booking flights over security concerns
United Airlines has announced that it is restricting unauthorized third-party AI agents from booking flights and managing customer reservations, citing concerns about data security, account access and unintended transactions.
The November 11 decision comes as AI assistants become increasingly capable of making purchases and completing tasks on behalf of consumers.
United warned that sharing account credentials with AI agents could expose sensitive personal information and lead to unauthorized purchases, account modifications or the use of MileagePlus miles.
The airline also raised regulatory concerns, noting that automated check-ins could bypass FAA and DOT requirements for passengers to personally acknowledge that their baggage does not contain prohibited items.
Kevin Surace, CEO, TokenCore
“I think United has two concerns here, and only one is being discussed publicly. The first is commercial. AI agents will become exceptionally good at finding the lowest fares, comparing every possible alternative, and making purchasing decisions entirely in the customer’s interest. Airlines have spent decades optimizing pricing, merchandising, and upselling. An intelligent agent working exclusively for the traveler will disrupt that model considerably.
“The second concern is legitimate liability. What happens when an agent books the wrong flight, spends someone’s MileagePlus miles, or purchases a ticket the customer insists they never authorized? Who pays when the customer disputes the charge? These are real issues, as are security and regulatory compliance. But they are solvable through verified identity, explicit customer authorization for financial transactions, spending limits, and auditable consent.
“Ultimately, I believe United will lose this battle. AI agents will increasingly act on our behalf, just as executive assistants have done for decades. We don’t prohibit executive assistants from booking flights because they might make mistakes. We establish authorization and accountability. The same will happen with AI agents. United’s decision may make sense today, but five years from now, preventing customers from using their own intelligent assistants to book travel will seem as antiquated as requiring everyone to call an airline ticket office.”
ㅤ
John Strand, Owner, Black Hills Information Security
“I think United’s caution makes sense. When you let an AI agent act on a customer’s behalf, you’re giving it access to parts of the site that are only available after login. A malicious agent could look for weaknesses in those areas, where it may be able to do more damage. From the outside, it’s hard to tell whether United is acting out of general caution or has specific concerns about what agents could do once authenticated.
“I also understand the concern about having passengers personally acknowledge that their bags don’t contain prohibited items. That’s a declaration you’d want the passenger to make directly. I expect we’ll see more of this from SaaS vendors. They’ll increasingly treat their authenticated services as walled gardens and decide which agents, if any, they’re willing to let in.”
ㅤIt’s about time that someone did this as AI agents do not have the required controls present. Hopefully others follow this lead.
Share this:
Like this:
Related
This entry was posted on November 13, 2026 at 7:30 am and is filed under Commentary with tags AI, United Airlines. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.