Exclusive-Meta testing a ’human concierge’ for its new personal AI agent Muse

New York, September 22 - Meta has been piloting a "human concierge" service for its personal AI assistant, Muse, according to internal company documents obtained by Reuters. This involves contracting human agents to manage some phone calls made through the digital assistant. The company announced the test to employees last week, following the public launch of a phone calling feature for Muse.

Some employees have expressed privacy concerns, warning that sensitive information could unintentionally be shared with contractors in call centers. A Meta spokesperson stated that the response from employees has been overwhelmingly positive, and the test aims to gather feedback to implement safety and privacy protections before a public rollout.

Daniel Roberts, a Meta spokesperson, added that Meta is working with merchants to improve the calling feature and will only release it when it is ready with proper disclosures. The Muse assistant has been a hit, with over 2.5 million downloads in two weeks, according to Sensor Tower data. CEO Mark Zuckerberg's vision for Muse is to provide personal superintelligence to millions of Meta users by performing tasks such as sending emails, shopping, and booking travel.

The assistant is built with safety protections, including secure "virtual machines" for each session and separate secure storage for sensitive data like passwords. The phone calling feature allows Muse users to dial US business numbers and handle tasks like booking appointments or checking product availability. After the call, the assistant provides a summary and transcript of the interaction within the Muse app.

Meta began testing the phone calling capability in August and gradually rolled it out to employees in the days following the app's debut. Half of Meta's employees were initially given access to the human concierge feature, as announced in a company post. Employees who preferred not to participate could join an opt-out group.