Meta tests human concierge feature for Muse AI assistant

by@Reuters

Summary

Meta is currently testing a "human concierge" feature for its new personal AI agent, Muse, which allows human contractors to manage some phone calls made through the digital assistant. This trial, which began after the recent public launch of Muse's phone calling capabilities, has raised privacy concerns among employees about the potential mishandling of sensitive information by call center contractors. In response to these concerns, a vice president at Meta acknowledged the oversight and noted that the feature has been temporarily rolled back. As part of its strategy to improve user experience and maintain privacy, Meta aims to refine Muse's capabilities before a wider release, a critical step given that the app has already gained over 2.5 million downloads and is central to CEO Mark Zuckerberg's vision for enhanced personal superintelligence.

Tokens

$META

Analysis

Meta: Meta Platforms develops social media platforms including Facebook and Instagram along with advanced AI technologies. The company is currently testing a human concierge system to support its personal AI agent Muse during phone calls to businesses. A vice president in its SuperIntelligence Labs unit addressed internal feedback on the test rollout. Muse: Muse is Meta's personal AI assistant capable of autonomously handling tasks such as sending emails, shopping, booking travel, and placing phone calls. The agent is being tested with a human backup feature for calls to improve reliability before wider public release. It was publicly launched recently with emphasis on secure virtual machines and data protections. Katie Paul: Katie Paul is a Reuters journalist based in New York who reported the exclusive details on Meta's internal testing of the Muse AI agent. Her coverage drew from company posts describing the human concierge experiment and related employee privacy concerns. Daniel Roberts: Daniel Roberts serves as a spokesperson for Meta and has addressed questions regarding the Muse human concierge test. He noted that employee responses have been positive and stressed the goal of adding safety and privacy measures prior to any public rollout of the calling feature. AI Reliability: Tech companies often incorporate human oversight into AI agent operations to handle tasks like phone calls where full automation faces user reluctance or lower success rates. Product Launch: Meta is iterating on safety features for Muse prior to broader release of its phone calling capabilities to merchants and users. Privacy Concerns: Internal testing of AI features involving external contractors has prompted discussions within Meta about proper disclosures and protections for sensitive user information.

Categories

techaiai_agentsvirtuals
View Original Tweet