Fast-changing AI rules: Companies are banning some agents; read the list of limits and exceptions

Why companies are drawing lines

AI agents have grown more popular since Meta released Muse in September, alongside tools such as Instinct. ChatGPT and Claude have also added agent features. Companies have pointed to privacy and security concerns. There has also been speculation that websites could lose traffic if people rely on agents instead of visiting directly to browse, shop or make bookings.

Retail and booking restrictions

Booking.com bars automated assistants from making bookings or reservations without express permission. Its terms also prohibit AI-powered assistants from accessing the platform for any reason without prior permission.

Some services approve selected agents

Resy restricts which AI agents can access its platform. It supports discovery and booking through ChatGPT and Claude, while retaining the ability to block other AI activity. Target's terms allow approved agents to sign in, build carts, place orders and start returns, with customers' permission. Target says it can change the approved list or restrict access at any time.

United and Yelp set conditions

United Airlines announced Wednesday that it was restricting third-party agent access, citing security and regulatory risks, along with the possibility of unintended purchases. Its policy bars unauthorized AI agents from searching, booking, buying or managing trips on its platforms.

Yelp prohibits non-human traffic without express permission. Its paid data-licensing program may let agents access data and take certain actions through approved integrations, including one that allows ChatGPT users to make reservations.