Overview
- Meta launched Muse as a consumer personal AI agent that can act across email, calendars and shopping and opened a developer connector API after its Sept. 8 launch.
- Amazon says Muse appears to capture and store customer credentials, does not identify itself when it browses, and therefore violates Amazon’s Conditions of Use, so the retailer has blocked the agent from shopping on Amazon.com.
- Meta’s design lets Muse use public APIs with user-provided credentials or operate “through a browser” inside a per‑user secure virtual machine when no API exists, a method that Amazon and other platforms have questioned.
- The dispute spotlights a core tension for agents: they need deep access to accounts to perform multi‑step tasks, which concentrates privacy and security risk and raises legal and contract disputes over platform control.
- Analysts warn the trust gap could limit paid subscriptions: Wall Street models show Muse would need very large conversion to reach tens of billions in revenue, so platform cooperation and user trust are key to the business case.