Meta's Muse AI assistant profiles users' friends and family, raising privacy concerns

Meta's new personal assistant, Muse, has become a viral hit with millions of downloads. It constructs detailed personal profiles of users' friends, family, and social connections by compiling data from connected accounts and public information.

Users connect Muse to services like bank accounts, messages, or health data so it can complete tasks. Its core functionality relies on creating a dedicated page for every person in a user's life. This hourly profiling process compiles data on family, partners, friends, colleagues, collaborators, and people the user follows.

Muse uses structured text files, termed 'memory,' to collect information. A profile may start sparse but fills over time with detailed sections. Meta's instructions direct Muse to use only available 'evidence,' stating that invented details are worse than an empty page.

How Muse works and what controls users have

Muse operates via an isolated virtual machine for each user, storing their data and context. This VM is inaccessible to other agents. Users can wipe memories or disconnect external services at any time. Meta states Muse seeks human confirmation before actions like sending email and maintains an audit log where users can see all agent activity and future plans.

Meta spokesperson Daniel Roberts told WIRED that Muse gathers context from public information and user-shared data. The system remembers context, such as identifying an invoice sender as a previously hired plumber or recalling a spouse's flower preference.

Expert concerns over data collection and user behavior

Researchers extracted Muse's internal operating instructions via the chat interface. Independent AI safety researcher Karan Joshi shared them with WIRED. Meta stated it intended for these files to be accessible for transparency.

Experts warn the tool encourages extensive data sharing. Miranda Bogen, director of the Center for Democracy and Technology's AI Governance Lab, said Muse emphasizes relationships and personal contacts more than rivals. These tools solicit users to plug in emails, calendars, and financial institutions, leading to a ballooning of what companies know. Carissa Véliz, an associate professor at Oxford's Institute for Ethics in AI, highlighted the imbalance. "We are giving AI systems much more information about us than we are getting information from them," she said.

Broader context: Meta's AI push and regulatory pressure

Muse is described as one of the most successful consumer AI launches to date, with download numbers continuing to climb despite privacy concerns. Meta generates the vast majority of its revenue from advertising, which depends on detailed user profiles. The company has billions of users across Facebook, Instagram, and WhatsApp and has faced ongoing regulatory challenges in Europe and scrutiny since the Cambridge Analytica scandal.

European regulators have begun scrutinizing AI data practices, and U.S. lawmakers are increasingly vocal about the need for AI oversight. Meta's privacy practices with Muse could become a test case for regulatory action.