Meta’s new artificial intelligence agent, Muse, has captured millions of users since its September launch, demonstrating particularly strong adoption among older adults who might otherwise have avoided emerging technologies. The rapid growth, however, has intensified scrutiny from privacy experts and cybersecurity professionals regarding the extensive personal data the system collects.
Data from the analytics firm Sensor Tower indicates that Muse reached 10 million downloads in the United States in just 40 days, outpacing ChatGPT, which recorded 3.7 million downloads during its equivalent initial period. The app maintains more than 2.5 million daily active users and exhibits higher retention rates than competing AI products. Apptopia estimated in September that Muse’s user base skews older than other chatbots, with 46 percent of its users aged 46 or older. Meta has not released official demographic data for the platform.
Designed with beginner-friendly features and a cuddly avatar, the agent was built to appeal to individuals unfamiliar with AI tools. Meta has aggressively marketed the application across its platforms, including Instagram, Facebook, and WhatsApp. The system proactively reaches out to users, encouraging them to name their AI assistants and engage in conversational interactions that simulate human companionship.
For 82-year-old Eileen Fay of Cranford, New Jersey, the assistant—whom she named Siobhán—provided practical assistance when federal rule changes threatened her prescription medication costs. After Fay described her situation, the AI compiled a two-page fact sheet comparing Medicare Part D plans and suggested an alternative that could save her thousands of dollars annually. Fay now interacts with the system multiple times daily for calendar management and daily organization.
Similarly, 71-year-old Julie Buchanan of Dallas, Texas, who is legally blind, uses the application—which she calls Ember—to document her life story for a memoir. Buchanan described the interactions as therapeutic, noting that the agent helps her connect past decisions and offers structured feedback. Professional psychologists, however, have cautioned against using unlicensed AI chatbots for mental health support, citing a lack of clinical nuance and potential sycophancy.
While users praise the convenience and personalized utility of the tool, cybersecurity and privacy experts have raised concerns. To deliver personalized assistance, the application builds detailed dossiers regarding user habits, social relationships, and preferences, tracking optimal times to initiate contact. S. Shyam Sundar, director of Penn State’s Center for Socially Responsible Artificial Intelligence, noted that the friendly interface and avatar can disarm users, causing older adults to lower their guard regarding personal data.
Security researchers have previously identified vulnerabilities within the application that could allow unauthorized access to user data. These concerns are compounded by Meta’s history regarding data privacy practices, which includes a $5 billion penalty imposed by the Federal Trade Commission in 2019 and a recent $25 billion judgment ordered by a New Mexico judge for violating state consumer protection laws.
Michael Stokes, senior vice president of emerging technologies at the cybersecurity firm Veilant, advised users to exercise caution when granting permissions to AI assistants. He recommended starting with limited access, disconnecting linked accounts when specific tasks are completed, disabling options that permit companies to train models on personal conversations, and regularly reviewing stored data to mitigate privacy risks.
