The Agentic Post
Breaking
Perplexity vs ChatGPT Search vs Google AI Mode: Which AI Search Is Best in 2026?  ·  What Are Vision-Language-Action (VLA) Models? How Robots Learn to Follow Instructions  ·  How to Translate Documents With AI Accurately (and When Not To)  ·  MCP vs A2A: The Two AI Agent Protocols Explained in Plain English  ·  Is Grok Accurate for News? What Studies Show About AI Search Citations  ·  Is AI a Bubble? 7 Warning Signs Investors Are Watching in 2026  ·  
Home/AI Agents/Computer-Use Agents
United Airlines Bans AI Agents From Booking Flights: What Travelers Need to Know

United Airlines Bans AI Agents From Booking Flights: What Travelers Need to Know

Computer-Use Agents

United Airlines changed its rules to stop unauthorized AI agents from buying tickets or managing bookings. What is banned, what still works, why United did it, and how other airlines differ.

United Airlines has banned unauthorized AI agents from buying its tickets, changing reservations or accessing MileagePlus accounts, even when the passenger approves each step. Announced on November 11, 2026, the change lets United cancel bookings made by such tools without notice. You can still use AI to research and compare United flights; you just have to complete the purchase yourself.

What exactly did United ban?

United revised its Contract of Carriage and its rules for travel agencies to bar “unauthorized third-party artificial intelligence agents” from transactions, Law Commentary reports. Blocked actions include:

  • Buying tickets and extras such as seats and upgrades
  • Changing or managing reservations
  • Using or applying MileagePlus miles
  • Logging into or changing customer accounts

Travel agencies and booking sites now need United’s express written consent before using AI agents to search, display, book or service its flights, or they risk losing the right to issue United tickets.

What can I still do with AI?

Plenty. You can ask an assistant to find routes, compare prices and suggest the best times to fly, and fare search on sites like Google Flights, Kayak and Hopper continues, Gadget Review notes. United says bookings made by AI agents before the announcement will be honoured. For tips, see how to plan a trip with AI.

Why is United blocking AI agents?

  • Account security: United says sharing logins with AI assistants can expose sensitive information and lead to unintended purchases or use of miles
  • Safety rules: FAA and DOT rules require passengers to personally confirm their bags contain no prohibited items like explosives; an AI clicking “agree” raises questions about who made that promise
  • Costs: automated searches that never turn into purchases add technology and distribution costs
  • Customer relationship: analysts say United does not want another powerful middleman, like Google, between it and its customers

Are other airlines doing the same?

No, approaches differ. Delta allows certain AI assistants to complete bookings when passenger credentials can be authenticated, with safeguards. American says some AI tools can help with purchases while it monitors automated activity. Southwest offers its inventory through ChatGPT but completes purchases on its own site. Other big companies, including Amazon and some banks, have also limited AI agents; see the Perplexity and Amazon court case.

What does this mean for AI agents?

It shows the next big fight around computer-use agents: businesses want control over who acts on their sites. Industry efforts like the Personal Agent Protocol aim to create approved, verifiable ways for agents to act for customers. United’s use of the word “unauthorized” suggests it could approve partners later, but none has been announced.

Should I give an AI agent my airline login?

Not for United, and be cautious everywhere. Never share passwords, card details or loyalty logins with an assistant that is not officially supported. Let the agent research, then book yourself. Our agent risk guide explains why.

Up Next
Anthropic IPO: $70 to $78 Price Range, Up to $1.4 Trillion Valuation, as Fed Official Warns of a Correction

Anthropic IPO: $70 to $78 Price Range, Up to $1.4 Trillion Valuation, as Fed Official Warns of a Correction

Claude

Anthropic set its IPO at $70 to $78 a share under the ticker ANT, valuing it at up to $1.4 trillion, while Fed Governor Michelle Bowman warned AI could trigger a large market correction.

Anthropic has set the price range for its stock market debut at 70 to 78 dollars a share, offering 825 million shares on Nasdaq under the ticker ANT. That could raise up to about 64 billion dollars and value the Claude maker at up to roughly 1.4 trillion dollars fully diluted. Pricing is reportedly set for November 18, with trading from November 19, just as a Federal Reserve official warned that disappointing AI returns could cause a “large correction” in stocks.

What are the key Anthropic IPO numbers?

From the November 10 amended filing, summarised by FinanceFeeds and Fortune:

  • Price range: 70 to 78 dollars per share
  • Shares: 825 million Class A shares, almost all sold by Anthropic, plus an option for banks to buy 123.75 million more
  • Raise: about 57.8 to 64.4 billion dollars
  • Strategic investors: Google and Nvidia (about 10 billion dollars each), AMD, Broadcom, Micron, Microsoft and TSMC will buy 26 billion dollars of non-voting shares at the IPO price
  • Valuation: roughly 0.9 to 1.0 trillion dollars on basic shares; reported at up to 1.3 to 1.4 trillion fully diluted
  • Exchange and ticker: Nasdaq Global Select Market, ANT
  • Lead banks: Goldman Sachs and Morgan Stanley, in a syndicate of 22 banks
  • Lock-ups: up to 180 days for most existing holders and up to 365 days for co-founders and some large holders

Why is the valuation lower than expected?

Some reports a few weeks ago suggested around 2 trillion dollars. Since then AI stocks have slumped, including a 2 trillion dollar one-day selloff after the midterms; see our report on the Nasdaq correction. A lower range makes it easier to sell 825 million shares into a nervous market. Anthropic’s business is still growing fast: Fortune reports an annualised revenue run rate of about 81 billion dollars at the end of October, up from about 71 billion in late September.

What did the Fed official warn about?

Fed Governor Michelle Bowman warned of a “large correction” in stock prices if AI investments fail to deliver expected returns, Semafor reported via Yahoo Finance. She pointed to about 800 billion dollars of AI-related debt, off-balance-sheet liabilities and circular financing, and said a shock “would likely slow the AI buildout” and could soften the economy. A Nomura strategist said any wobble around the IPO “might mean a selloff of the AI theme.”

Can ordinary investors buy Anthropic shares?

Before trading starts, only through brokers that receive an allocation, reportedly including Robinhood and SoFi among a handful of retail platforms. Most people will be able to buy once ANT trades on Nasdaq. First-day prices for hot listings can swing wildly. Read how to invest in AI stocks first. This is not investment advice.

What does the filing say about AI safety?

The filing repeats a striking warning we covered in Anthropic’s IPO filing on AI accelerating research: future models “may imminently meet” its threshold for dramatically accelerated progress attributable to AI, which it says “could also pose a considerable safety risk.” For the basics, see what self-improving AI is.

What happens next?

Watch the final price on November 18 against the range, first-day trading on November 19, and whether the debut lifts or sinks other AI stocks. The listing comes days after a Claude login outage, and investors will also be weighing Anthropic’s 895 billion dollars in computing commitments.