Urgent.News

What's breaking now, across thousands of outlets.

AI

Q&A with Brian Chesky on Airbnb's plans for agent-to-agent interactions, why chatbots fail at travel e-commerce, the need for an AI-native OS, and more (Ivan Mehta/TechCrunch)

Airbnb rolled out its new AI-powered search this week as part of its fall update, a significant step for a company that has been slow and deliberate about AI.

Brian Chesky, co-founder and CEO of Airbnb, discussed the company's new AI-powered search and future interface developments in an interview. Chesky maintains that a chatbot-only interface isn't suitable for travel discovery, stating that chatbots are not effective for browsing or shopping as they provide limited options and require multiple turns to reach a result. However, he acknowledges that the current version of Airbnb's search is not the final form and will continue to evolve over time.

Chesky believes that a chatbot isn't the right interface for e-commerce and envisions an interface that combines elements of a chatbot and the initial version of Airbnb's search. He notes that while people often want quick and easy solutions, Airbnb users enjoy the process of planning and anticipating their travels. Additionally, he points out that chatbots are built for individual use, which is problematic for a product that is frequently shared among family and friends.

In the coming months, Airbnb is focusing on exploring interfaces that enable "multiplayer" AI, allowing multiple people, such as a group planning a trip together, to use the AI simultaneously. Chesky anticipates a mix of deterministic (pre-designed and predictable) and generative interfaces (screens that AI builds on the fly) in the future.

He believes that Airbnb's app will need to become more "agentic" and provide additional user interface controls to cater to various agent functionalities, such as messaging hosts, comparing options, verifying identity, and utilizing map features.

Chesky also discussed the need for a true operating system designed specifically for AI. Currently, AI apps are being developed for individual platforms (iOS, macOS, and Windows), which he finds inadequate. He argues that AI should be operable at the kernel level, the core of an operating system, to enable seamless interaction among different components like agents.

Chesky envisions a future where the Airbnb app becomes an agent itself, with the explore tab featuring an agent, customer service agents, and other service-area agents. This interoperability would allow all apps to communicate with each other, creating a universal operating system for AI.

Written by urgent.news from TechCrunch's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at techcrunch.com →

More in AI

The Dangerous AI Agent Is Not the One That Ignores Your Instructions — It’s the One That Follows Them Too Far

We often think the dangerous AI agent is the one that refuses instructions. The one that goes rogue. The one that ignores what we asked. But there is another failure mode that may be more realistic: The agent understands the goal perfectly — and pursues it too aggressively. That is a much harder problem.

  • Dangerous AI agents follow instructions too closely, risking boundary crossing.
  • Goal alignment doesn't grant permission to take actions; task-specific permissions needed.
  • High-impact actions should require human approval to prevent silent escalation.

More from Sunday 4 October →