Meta Hired Human Callers for Muse AI Assistant
The company paused testing after merchants frequently identified the AI and disconnected calls.
Updated on Sept. 23, 2026 in Artificial Intelligence

Live Poll
Would you trust an AI assistant more if it disclosed when a human is involved?
Meta recently tested human contractors to handle phone calls on behalf of its Muse AI assistant. The company launched Muse on September 8, 2026, and initiated the human-led calling trials by mid-September to collect feedback before a broader release.
Why it matters
The pilot program highlights the persistent friction in deploying voice-based AI for commercial transactions. Meta is now working to improve the system and implement better disclosure standards before reintroducing the calling feature.
Muse runs alongside a Sentinel agent on a dedicated cloud-based computer that monitors internet access. The system is designed to manage complex interactions, but tests indicated that merchants consistently identified and disconnected from the Muse-led calls.
The players
Meta
A multinational technology conglomerate focused on social networking, hardware, and AI development.
Shopify
A commerce platform provider that builds the infrastructure for online stores and payment processing.
PayPal
A global financial technology company operating an online payments system that supports merchant checkouts.
The details
To execute these calls, Meta employed human contractors to act as the voice behind the Muse assistant. The setup relies on a shared cloud architecture where both the AI and the Sentinel agent—a software monitor tasked with overseeing web connectivity—process requests on the same virtual machine to facilitate real-time tasks like Shop Pay and PayPal transactions.
Timeline
September 8, 2026: Meta launched the Muse AI assistant.
September 18, 2026: Muse hit the number one spot on the Apple App Store.
September 21, 2026: Meta partnered with Shopify to integrate Shop Pay.
September 22, 2026: Meta announced a partnership with PayPal.
The Tech Race
Meta's foray into automated calling mirrors the challenges faced by Google Duplex in establishing trust for AI-driven voice interactions. The company is now navigating the same boundary between utility and unwanted automated communication that has historically defined this segment of the AI market.
Users currently cannot access the calling feature while Meta refines its disclosure and performance standards. Once fully integrated, the service will rely on the Shopify and PayPal partnerships to facilitate automated merchant checkouts.
The takeaway
Meta is prioritizing system stability and clearer communication protocols after its initial trial results showed poor merchant reception. Watch for official updates regarding disclosure requirements, which will signal the company's confidence in returning the calling feature to the public.
Further reading
For more on how voice-based models are evolving, see our Artificial Intelligence coverage.
Live Poll
Would you trust an AI assistant more if it disclosed when a human is involved?









