Extracted system prompts show Meta's Muse compiles "a page for every person in the user's life", with facts, history, tips to improve relationships, and more

Meta's AI is now building detailed dossiers on your friends, family, and colleagues, offering advice on your relationships.

Meta's AI chatbot, Muse, has been found to compile detailed personal profiles for every individual in a user's life. Extracted system prompts reveal that Muse creates "pages" for friends, family, colleagues, and others, documenting facts, relationship history, and even offering advice on improving connections. This hourly process utilizes Muse's "memory" to gather and structure information, aiming to provide personalized suggestions. While AI chatbots commonly handle social information, Muse's approach is notable given Meta's extensive social network data. Researchers expressed concern over the depth of personal understanding Muse aims to achieve, describing it as "creepy." Meta states that Muse relies only on available "evidence" and aims to avoid inventing details, prioritizing accuracy. The company also emphasizes Muse's architecture, including dedicated virtual machines for user data, user control over memory, and a requirement for human confirmation before actions. Meta spokesperson Daniel Roberts stated Muse gathers context from public information and user-shared data to personalize assistance. Experts note Muse appears to place a greater emphasis on relationships and personal contacts compared to other AI assistants, which generally encourage users to share more data.

Meta's Muse AI is architected to build comprehensive relationship profiles by systematically gathering and analyzing user-provided data and publicly available information. This deep dive into personal connections, including relationship history, shared interests, and even potential friction points, represents a significant push towards AI agents that not only understand individual users but also their social ecosystems. The emphasis on "strengthening" relationships suggests a move beyond simple assistance to active social management, potentially influencing user interactions and interpersonal dynamics.

This development signals a new frontier in AI's integration into personal lives, moving beyond task completion to influencing social behavior and relational depth. The extensive data collection and analysis capabilities of Muse raise profound questions about privacy, data ownership, and the ethical implications of AI actively participating in human relationships. As AI agents become more sophisticated in understanding and predicting social dynamics, users will face increasing decisions about how much of their social lives they are willing to entrust to these systems, and what benefits justify the inherent privacy trade-offs.