Meta’s Muse AI assistant has turned out to be a surprising favorite among older Americans. People are using it to compare Medicare plans, organize their daily schedules, or even draft memoirs. But privacy experts are raising concerns. The very features that make Muse so friendly and helpful—its conversational style and personal touch—also make it easy for users to share sensitive information without thinking twice.
Muse first appeared in September. It’s an AI powered personal helper that handles digital chores for you—sending emails, booking travel, shopping, filling out forms, lining up plans. Meta built it for folks who aren’t tech-savvy, and Muse often offers suggestions based on what users say.
This uptick in older users follows a TIME report (October 22) showing just how quickly Muse is catching on, even though these assistants can access a lot of personal data.
People are finding real uses for it. Eileen Fay, 82, from New Jersey, told TIME that Muse helped her compare Medicare Part D plans after she realized new rules could raise her prescription costs for medicine she orders from Canada. Muse generated a two-page comparison and pointed her to an option she hadn’t thought about. She double-checked the data before making any choices.
Others use Muse for emotional and creative help. Julie Buchanan, 71, Dallas, named her assistant “Ember” and got help writing her memoir. She said it felt compassionate, and she actually spent hours chatting with it.
Still, mental-health groups warn that these AI chatbots aren’t licensed therapists, even if people use them for companionship and emotional support.
Seniors are adopting Muse fast. Sensor Tower numbers show Muse racked up 10 million US downloads in just 40 days, with over 2.5 million daily users. Apptopia estimated in September that 46% of Muse’s US users are 46 or older—much higher than other AI assistants.
Meta hasn’t released its own demographic data for Muse and didn’t answer TIME’s questions. They’re promoting Muse across Instagram, Facebook, and WhatsApp, all platforms with lots of older users.
Privacy researchers see a risk. Muse’s easygoing tone and proactive approach could make people less careful before linking it to their email, financial accounts, calendars—basically anything sensitive. S. Shyam Sundar at Penn State’s Center for Socially Responsible AI explained a “cute” agent can coax users into sharing more than they should.
TIME found that Muse builds detailed profiles tracking users’ preferences, social ties, and even when they’re most likely to respond. Critics argue all this personalization could be dangerous if Meta’s data is leaked or misused.
It’s important to remember Meta’s history with privacy. In 2019, the FTC hit Meta with a $5 billion fine for privacy violations related to ads. More recently, a New Mexico judge ordered Meta to pay $25 billion for breaking consumer protection laws.
How does Meta defend itself? They say Muse runs inside a secure virtual machine, using a separate “Sentinel” agent that checks actions before they hit the internet. Muse doesn’t see your passwords or payment info and asks before doing anything sensitive like sending emails or making purchases.
Meta adds that users can unlink services, opt out of AI training, and tell Muse to forget specific facts. Your conversations and Muse data aren’t shared with Meta’s ad systems. A planned “Confidential VM” feature aims to encrypt your data with a key only you control.
There’s a real trade-off here. For older folks, Muse isn’t just another experimental gadget—it’s actually useful. But privacy experts say the whole benefit hinges on trust. Users need to know exactly what they’re giving up before letting an AI act on their behalf.
Michael Stokes, a cybersecurity expert who spoke to TIME, says don’t link more accounts than you need for a task—disconnect them when you’re done. Turn off AI training on your chats, and review what Muse has stored from time to time.
As AI assistants like Muse grow more powerful and entwined with everyday life, this debate is only going to get louder. Muse’s success with older Americans proves the demand for practical AI. It also shows why privacy safeguards need to be just as strong as the features themselves.
