Amazon blocked Meta's AI assistant, Muse, from completing purchases on Amazon.com over the weekend. TechCrunch's Russell Brandom reported on September 21 that users who tried to route a purchase through Muse got an error message instead, telling them that "continued access by an unauthorized AI agent violates Amazon's Conditions of Use, to which our customers have agreed." Brandom notes Amazon runs its own competing AI stack through AWS, and that agent-driven shopping has a liability problem baked in: if an agent orders the wrong item, Amazon is the one left cleaning up the dispute with the customer and the vendor, not Meta. Muse's error rate is already fairly low. Amazon blocked it anyway.
The Meta-versus-Amazon rivalry is the easy read here, but it's not really what caught my attention. A retailer just drew a line about who's allowed to shop on its own site, and the line wasn't drawn around a person. It was drawn around a category. Agents acting on someone's behalf, trustworthy or not, are now unauthorized by default until proven otherwise. That's a policy decision wearing a terms-of-use notice as a disguise, and I doubt Amazon will be the last company to make it.
For Vunoh, this is worth sitting with from both directions. If we're building or relying on tools that browse, click, or transact on our behalf, we should assume that some of the platforms we depend on will simply say no, the way Amazon just did to Meta, and plan around that rather than being surprised by it. And if we're ever on the other side, deciding whether to let outside agents interact with our own systems, we should be deliberate about it rather than defaulting either to blanket trust or a blanket block. Amazon has made its choice. We haven't had to make ours yet, but it's coming.
REFERENCES
Meta's AI agent has been blocked from using Amazon.com — TechCrunch, September 21, 2026: https://techcrunch.com/2026/09/21/metas-ai-agent-has-been-blocked-from-using-amazon-com/