Meta’s AI personal assistant Muse, released last month, continuously builds detailed dossiers on users and their contacts by accessing chats, emails, and social media interactions, according to internal instructions reviewed by Time Magazine and reported by medianama.com. Muse updates these dossiers every hour, tracking relationships, goals, and behavioural patterns to assist users with tasks such as grocery shopping and subscription management.
Muse’s internal guidelines direct it to maintain comprehensive records of users and people they mention, including social relationships, shared interests, disputes, and alliances. The AI also adapts its communication style by identifying inside jokes and shared context to strengthen user relationships. Importantly, Muse collects information not only on users but also on non-users who appear in conversations, expanding its network of data beyond direct users, medianama.com reported.
This approach marks a significant development in AI personal assistants by creating continuously updated profiles that infer users’ goals, even those unspoken, through nightly analyses of daily conversations. By mapping social connections and behavioural cues, Muse aims to deliver personalized assistance. The dossier-building capability raises questions about privacy and data use, given its extensive reach into personal communications and relationships.
Meta’s Muse AI began rolling out last month and is positioned as an assistant integrated with email and Instagram. The dossier updates occur every hour, with a nightly review of conversations to refine its understanding of users’ needs and social dynamics, according to medianama.com.