According to 404 Media, Meta is testing an AI agent called Muse that is designed to make calls. However, during the current test phase, the calls are actually being placed by human workers in a call center, not by the AI itself.
The report includes an internal comment expressing concern that this approach could backfire. The commenter worried that if it became known that humans were making the calls, it could generate negative press suggesting Meta's AI is not capable enough to operate without human backup.
This reveals a broader tension in the AI industry: companies want to showcase autonomous agents, but often rely on human fallbacks to ensure quality and avoid embarrassment. The source does not say when or whether Muse will eventually operate without human assistance.