Meta initiated a test involving human callers for its Muse AI assistant after merchants frequently disconnected upon identifying the caller as an automated system. The company informed employees about this trial last week, shortly after introducing phone-calling capabilities to the platform, though the test has since been paused.
Internal feedback raised concerns among staff regarding potential data privacy risks. Employees warned that involving human contractors could lead to the exposure of sensitive user information and undermine the security protocols previously highlighted by Meta during the product launch.
Meta spokesperson Daniel Roberts stated that the trial was intended to gather feedback before a public release, noting that internal staff response was positive. The company plans to continue refining the calling feature and will implement it only after ensuring proper disclosures are in place.
Muse, launched on September 8, functions as a personal AI assistant running on dedicated cloud hardware with a Sentinel security agent. The application reached the top position on the Apple App Store by September 18. Meta is actively expanding payment integrations for the assistant.
The platform initially launched with Stripe Link support. Recent updates include partnerships with Shopify to enable Shop Pay functionality and with PayPal to allow users to complete transactions at participating merchants using their PayPal accounts.
