On Sunday night, users attempting to buy items on Amazon.com through Meta’s AI assistant Muse began receiving an error message. As first noted by GeekWire, the message read: “Continued access by an unauthorized AI agent violates Amazon’s Conditions of Use, to which our customers have agreed.”
In practice, that means Muse is not permitted to act as a shopper on Amazon, and anyone trying to place an order via the AI agent will need to purchase the item elsewhere.
Reasons behind the block
The move can be seen as part of the broader dynamic between large tech companies: Amazon has its own foundation models and a widely used inference platform, so it is under no obligation to grant marketplace access to third-party agents like Muse. Beyond platform control, there are practical concerns:
- If an agent such as Muse places an incorrect order, Amazon would be responsible for addressing the fallout with both the customer and the seller.
- Although the reporting notes Muse has a relatively low hallucination rate compared to many AI models, it is not immune to errors.
Those operational risks help explain why Amazon might restrict agent-driven commerce for now or wait for further model iterations before allowing such interactions.
Implications
As it stands, the block prevents Muse from completing purchases on Amazon. The incident highlights how platforms enforce their terms of use and underscores the practical and liability questions that arise as companies experiment with automated, agentic shopping.
GeekWire was the first outlet to spot the error message; the source did not report additional official comments from Amazon or Meta.
