Amazon has blocked Meta’s new Muse AI agent from shopping on Amazon.com on behalf of users, escalating a broader battle over whether autonomous assistants should be allowed to navigate online stores without the retailer’s permission.
Customers attempting to use Muse on Amazon are now shown a message stating that continued access by an unauthorized AI agent violates Amazon’s Conditions of Use. Amazon says Meta did not tell the company that Muse would access its marketplace and declined a request to remove Amazon voluntarily from the agent’s supported shopping experience.
Muse, launched by Meta on September 8, is designed to perform multi-step tasks including managing emails and calendars, researching products and completing purchases. When a service lacks a dedicated API, Meta says the agent can operate through a browser much as a user would.
That mechanism sits at the center of the dispute. Amazon says Muse does not identify itself while navigating the site and can interact with account pages, order histories and credentials, creating potential privacy and security risks. Meta has previously said Muse cannot see users’ passwords or payment details. Credentials entered by customers are placed into secure storage so the agent can use them without directly accessing their contents.
The conflict is also commercial. AI shopping agents could shift product discovery away from Amazon’s own interface, where sponsored listings and advertising generated more than $68 billion for the company last year. An agent could instead compare products and prices according to a user’s instructions before going directly to checkout.
Amazon has already moved against other third-party agents, including products from Perplexity, Google and OpenAI, while developing its own agentic shopping capabilities.