Meta’s Muse has been shut out of Amazon. Shoppers who send the assistant to the retailer meet a warning telling them that “continued access by an unauthorized AI agent” breaches Amazon’s conditions of use, and the company confirms it never granted that access – it has asked Meta to remove its marketplace from the agent.
Amazon set out its position in a statement. Applications that offer to buy on a customer’s behalf for a different business, it argues, should work openly and accept a service provider’s decision on whether to take part. On Amazon’s reading, agentic apps such as Muse carry those same obligations.
Data handling sits behind the technical accusations. Amazon says the assistant scrapes account data, and appears to capture and store customer credentials – behavior it links to questions about sensitive information. The agent takes no step to identify itself as one. A separate argument concerns personalisation, since an agent that ignores the signals behind the storefront may surface products unrelated to a shopper’s history. Transparency, an opt-out for retailers and a mutual exchange of value are what Amazon is asking for.
Muse users were the first to spot it, posting about the block on social media before Meta commented. Meta did not immediately respond to a request for comment.
The standoff sets the ground rules for agentic shopping. An assistant can only buy on someone’s behalf where a retailer agrees to be bought from, and Amazon has shown that the largest of them will defend its storefront rather than quietly absorb the traffic.
