Third-party call center staff are managing some phone interactions for the Muse AI agent based on internal posts seen by Reuters. Meta launched this internal experiment shortly after the release of the calling feature to develop safety and privacy protections before the tool is released to a wider audience.

Staff at the company raised privacy warnings that users might share sensitive personal information without knowing a human operator is involved in the conversation. Meta will include formal disclosures for users if the phone-calling feature is launched publicly, though the company stated the test is currently limited to internal use.