Technology | October 9, 2026
Meta’s personal AI assistant, Muse, is drawing attention over how much personal information it can analyse and retain. A report by TIME, cited by Business Today, says the assistant can build detailed profiles from conversations, messages and emails it accesses, including users’ preferences, social relationships and behavioural patterns.
How Muse Builds Personal Profiles
According to the report, Muse can update user profiles hourly and analyse conversations each night. Its profiles may include people users interact with, shared interests, past disagreements and communication preferences. It can also infer goals and identify approaches that may make its suggestions more effective.
Information About Other People May Also Be Included
The reported profiling may extend to people mentioned in conversations, even if those individuals have never used Muse. This raises questions about how AI assistants process information about third parties and how much control those people have over data shared about them.
What Meta Says About Privacy
Meta says Muse remembers information users choose to share to make its assistance more useful. The company also states that conversations and virtual-machine data are not shared with Meta’s advertising systems. However, Meta says sanitised interaction data may be used to train and improve its AI models, and users can opt out through the settings.
Users Have Data Controls, but Questions Remain
Users can inspect, edit and download files stored in their Muse environment. They can also ask the assistant to forget information. However, TIME’s analysis reportedly found that removing a memory may not automatically delete the original conversation from which it was learned. Meta has said it plans to introduce additional encryption protections later in 2026.
The report highlights a broader challenge for AI assistants: balancing personalisation and convenience with transparency, data minimisation and meaningful user control.
