Amazon Blocks Meta's Muse AI Shopping Agent

Amazon has blocked Meta’s new personal AI assistant, Muse, from shopping on its platform. Users attempting to use the agent to purchase items on Amazon.com now encounter a warning stating that unauthorized AI agents violate the site’s conditions of use. This move marks a significant escalation in the ongoing tension between the two tech giants over how artificial intelligence should interact with e-commerce ecosystems.

The block comes despite the companies being business partners, with Amazon products already available on Facebook and Instagram. Amazon claims it was not informed in advance that Muse would access its store and that the agent does not identify itself when browsing. According to GeekWire, Amazon argues that Muse captures customer credentials and scrapes account data without proper authorization, posing security and privacy risks.

Security concerns drive the block

Amazon’s primary objection is that third-party agents must operate openly and respect the decisions of service providers. The company states that Muse appears to store customer credentials and access sensitive account pages, including order history, without the retailer's knowledge. Amazon compares this to unauthorized access, arguing that legitimate intermediaries, such as travel agencies or food delivery apps, work in partnership with businesses rather than bypassing them.

Meta launched Muse in September as a tool to handle multi-step tasks, including shopping, by connecting to various services. The agent runs on a secure virtual machine and can use public APIs or browser-based access if APIs are unavailable. However, Amazon insists that this mechanism allows Muse to move through customer accounts and process transactions without the retailer’s explicit consent, undermining the safety and reliability of the customer experience.

Legal battles over AI access

This standoff is part of a broader legal and regulatory debate over who controls the online shopping experience when AI agents act on behalf of consumers. Amazon has previously sued Perplexity and moved to block agents from Google and OpenAI, citing similar concerns about unauthorized access. While a recent Ninth Circuit ruling in the Perplexity case limited Amazon’s ability to claim hacking, it left open avenues for enforcement through terms of service and contract law.

The current message shown to Muse users cites Amazon’s Conditions of Use rather than accusations of hacking, reflecting this legal strategy. Amazon declined to comment on whether it would pursue legal action against Meta in this specific instance. The company says it is in direct conversation with Meta to resolve the issue, emphasizing the need for transparent and authorized interactions between AI agents and retail platforms.

High stakes for ad revenue

The dispute carries significant financial implications for Amazon, which generated over $68 billion in advertising revenue last year. This business model relies on customers browsing its pages and engaging with sponsored products. If AI agents bypass the traditional browsing experience, Amazon risks losing control over where and how its ads are displayed, potentially diminishing the value of its advertising inventory.

Meta did not immediately respond to requests for comment regarding the block. The situation highlights a fundamental conflict between the emerging model of agentic AI, which prioritizes seamless task completion for users, and the established e-commerce model, which prioritizes platform control and data security. As AI agents become more common, the question of how they should interact with existing digital services remains unresolved and increasingly critical for both consumers and businesses.