Meta’s Muse draws older adults into personal AI—and raises privacy concerns

Meta built Muse to make AI approachable for beginners. TIME’s October 22 reporting shows older adults using it for Medicare comparisons, daily planning and memoir writing. But the same personalized attention that users value worries privacy experts, who warn that a friendly assistant can make sharing sensitive information feel less consequential.

From Medicare arithmetic to a life story

For Eileen Fay, 82, the appeal was concrete. Facing potentially higher medication costs, she asked Muse to compare Medicare Part D plans. It suggested two options, including an unfamiliar plan that she said could save her more than $6,000 a year. Fay told TIME she checked the information and found it accurate.

Comparing costs for nine medications would have been complex, Fay said; Muse finished in 10 minutes. She now speaks with the assistant, which she named Siobhán, several times a day to manage her calendar and to-do lists.

Julie Buchanan, a 71-year-old retiree who is legally blind, uses Muse to write her memoir. She told TIME the assistant, which she calls Ember, feels more compassionate than therapists she has seen. Mental health experts caution against using unlicensed chatbots as therapists.

I find I have to keep reminding myself that this is not a friend who’s getting to know me better.

Eileen Fay, speaking to TIME

An older audience, with limits to the data

Apptopia estimated in September that 46% of Muse’s U.S. users were 46 or older. That age band is broader than the retirees TIME interviewed. Meta has not released demographic figures and did not respond to TIME’s request for comment.

Sensor Tower figures cited by TIME put Muse at 10 million U.S. downloads within 40 days of its September launch and more than 2.5 million daily active users. Those are overall figures, not older-user counts.

Helpful attention requires personal information

Muse reaches out with suggestions and builds profiles of users’ wishes, relationships and messaging habits, including when they are likely to respond. After helping Fay prepare for her TIME interview, it followed up without prompting to ask how it went.

S. Shyam Sundar, director of Penn State’s Center for Socially Responsible Artificial Intelligence, warned TIME that a cute assistant can make older adults lower their guard about personal data. He has not published research specifically on Muse. TIME also cited The Information’s reporting of a vulnerability allowing an attacker to access user data.

Michael Stokes, a senior vice president at cybersecurity company Veilant, advised users to:

• Connect only accounts needed for a task, then disconnect them afterward.
• Turn off the option allowing Meta to train its AI on conversations.
• Periodically review what the assistant has saved about you.

His warning extends beyond account holders: shared information can reveal details about family and friends who never agreed to share them.