Meta’s Muse is being described as an effective AI assistant, though one that has raised eyebrows over how it handles personal data. Much of the concern stems from its new Mac app, which can access Messages, Calendar and Notes. Yet for all its capabilities, Muse appears unable to accurately explain how it actually works.
Jason Aten, a contributing editor at Inc Magazine, shared screenshots of an exchange in which the assistant asked him questions about a conversation he was having in Messages. The difficulty, according to Aten, was that he had not granted Muse access to his messages in the first place.
What Muse said about accessing messages
When asked how it knew the contents of his messages, Muse replied that it had seen the notification previews rather than his message history, adding that it had not been reading his texts. Pressed further on how it was receiving message previews, the assistant struggled to provide a clear response. It stated that it could not give the exact technical detail, explaining only that the paired Mac app exposes notifications as one of its capabilities and that these arrive through device sync.
The exchange highlighted a wider issue around transparency, as the assistant offered an explanation it could not fully substantiate.
Meta’s response to the privacy questions
David Singleton of Meta Superintelligence Labs responded to the post to clarify the situation. He outlined the permissions Muse requires to read messages, including granting the Mac app full disk access, and stressed that the features are opt-in.
Singleton said that Muse does not watch notifications on a Mac, but instead syncs data from Messages only after the user has specifically enabled access. He suggested the issue was not that Muse had improperly read the texts, but rather that the assistant did not understand how to describe its own functionality.
According to Singleton, when Muse said it synced device notifications, it was confused about how to explain the feature and gave an incorrect explanation. He apologised for the incorrect response and said the company is working to improve Muse’s understanding of its own internals so that it answers questions about how it functions more consistently.
The episode underscores a familiar characteristic of AI chatbots, which tend not to reveal their inner workings and instead offer answers that appear plausible. For users, the takeaway centres on the opt-in permissions and the full disk access that Muse’s Mac app requests before it can sync data from Messages.
Source
Image: theverge.com
