Amazon Blocks Meta’s Muse AI Agent From Shopping

Amazon has blocked Meta’s new Muse AI agent from browsing and making purchases on Amazon.com, after asking Meta directly to exclude the site from Muse’s shopping capabilities and getting no cooperation. Starting Sunday night, September 20, users trying to check out through Muse hit a popup instead, warning that unauthorized AI agent access violates Amazon’s Conditions of Use.

Here’s the part that matters more than the block itself. Amazon says Meta never asked permission before Muse went live and started shopping on its platform. The agent also doesn’t identify itself while browsing, according to Amazon, and appears to store customer login credentials, something Meta disputes.

Muse launched September 8 alongside Meta’s own Muse Spark model. It can browse the web, fill out forms, and complete purchases once a user approves the action. Meta says every session runs inside a dedicated secure virtual machine, with passwords and payment details kept out of the agent’s reach in protected storage.

This isn’t Amazon’s first fight over agentic shopping, and it won’t be the last. The company sued Perplexity over its Comet browser last year, and it has already restricted crawling access for bots from OpenAI, Google, and Anthropic. Amazon’s stated position is simple. Automated systems have to identify themselves, and stop accessing its services the moment Amazon asks.

Not everyone’s taking Amazon’s side on principle. Elon Musk questioned on X whether the restriction is even enforceable long term, pointing out that once agents operate through a user’s own browsing session, cookies and IP address included, Amazon has no real way to tell a human shopper from an AI acting on their behalf.

And Meta isn’t backing off either. On the same day the block escalated, Meta and Shopify announced a partnership letting Muse users shop and check out across Shopify stores directly through Shop Pay. Shopify’s CEO has said explicitly that Meta’s agent will be allowed to complete purchases on its storefronts, a stance that puts Shopify at odds with Amazon’s approach.

The bigger question underneath all this isn’t really about Muse. It’s about who owns the customer relationship once an AI agent is the one doing the browsing, comparing prices, and clicking buy. Amazon has an obvious financial reason to keep that relationship for itself, most of its revenue outside retail comes from advertising tied directly to how people shop on its site. If agents start making those decisions instead of humans scrolling through search results, that entire model gets harder to defend.

Banks are watching this fight too. Several major banks, including NatWest and Bank of America, have already flagged that AI shopping agents raise the risk of scams and financial fraud, mostly because nobody’s agreed yet on how an agent should prove it’s acting with a customer’s actual authorization.

This dispute connects to a pattern we’ve covered before. See our earlier piece on Amazon’s lawsuit against Perplexity getting the green light to proceed, which set the stage for exactly this kind of standoff. For more on Meta’s AI product expansion beyond shopping, check out our coverage of Meta’s native Mac app for conversational AI. And if you’re evaluating AI tools built specifically for online retail, our AI ecommerce assistant category covers what’s currently available.