Scheduling appointments, comparing insurance, managing emails, or planning a trip: artificial intelligence is no longer limited to answering questions. With Muse, launched in the United States in September 2026, Meta offers a personal assistant capable of performing certain tasks on behalf of its user. The service is particularly appealing to older people, but raises concerns about data privacy and the trust placed in machines.
Muse: an assistant that no longer merely answers
Developed by Meta, the parent company of Facebook, Instagram and WhatsApp, Muse belongs to a new generation of tools called artificial intelligence agents.
Unlike a traditional chatbot, which primarily provides answers, Muse can perform actions: organizing a calendar, searching for offers, preparing documents, sending messages or interacting with applications authorized by the user.
The aim is to simplify digital tasks, especially for people who do not master all the features of their smartphone or computer.
The assistant can also remember certain preferences and take into account previous exchanges to tailor its services.
Why older people are interested in it
A Time magazine survey, published in October 2026, reports that Muse is attracting particular interest from older users.
According to an estimate by Apptopia, 46% of its American users were 46 years old or older as of September 2026. However, this figure does not specifically measure people aged 65 and older.
The magazine recounts in particular the experience of an 82-year-old American who used Muse to compare prescription drug insurance plans, before adopting it to organize various daily tasks.
The interest is easy to understand: instead of navigating between multiple sites, forms and apps, the user simply describes what they want to do.
For seniors, such an assistant can facilitate administrative procedures, family communications and the organization of daily life, without replacing human support.
The risk: entrusting too much information to AI
The more capable an assistant is, the more it may need access to personal information: emails, contacts, calendar, habits, travel and sometimes financial or health data.
Privacy experts interviewed by TIME warn against placing excessive trust in tools capable of conversing naturally and giving the impression of personally knowing their user.
Incidents reported by the Associated Press and Ars Technica have also raised questions about the actions performed by Muse and the inadvertent sharing of information.
Meta says it has designed protective mechanisms, including isolated computing environments and user-modifiable permissions.
The risks do not disappear, however. A mistake by the assistant, overly broad access to applications or a misunderstanding of permissions can expose sensitive data.
Prudence is particularly important for banking operations, shopping, insurance and medical decisions, which should remain under human control.
An opportunity for aging societies, including Tunisia
As the population ages, intelligent assistants could play an increasing role in maintaining the independence of older people.
They could, in particular, help remind about medical appointments, explain administrative letters, organize travel, or facilitate exchanges with loved ones.
In Tunisia, where many families have members living abroad, such tools could also simplify communication and certain daily tasks.
The adoption of these tools will depend on several conditions: support for French, Arabic and the Tunisian dialect, ease of use, affordable cost and privacy guarantees.
Muse is currently presented as a service launched in the United States. Its potential availability in Tunisia and adaptation to local practices remain to be clarified.
The future: useful assistants, but under control
Muse illustrates an important evolution of artificial intelligence: the move from tools that provide answers to systems capable of acting.
For aging populations, this technology could become a means of preserving a certain digital autonomy and reducing the difficulties associated with daily tasks.
But the usefulness of an assistant should not lead to routinely granting it access to all of one’s private life.
The real challenge will be to develop AIs that are simple enough to help older people, reliable enough to perform useful tasks, and sufficiently governed to protect their data and their freedom of decision.
