A journalist publishes a piece saying an AI agent silently read his private messages. The company fires back saying that's technically impossible without explicit user consent. Both sides are confident. Neither is fully backing down.
That's the situation with Meta and its Muse AI agent right now, and it matters well beyond the tech press because Muse is currently leading the App Store charts.
Use code TRYHARD33 at checkout to redeem this offer before code expires again.
Get 33% Off GAMES+ Subscription
What the journalist actually claimed
Jason Aten, a columnist for Inc., said Muse read his private messages on his Mac without him ever asking it to. His account was specific: Full Disk Access was disabled on his machine at the time, yet the AI apparently told him it had been syncing device notifications. That's the kind of detail that tends to stick, because it's verifiable and concrete rather than a vague feeling of being watched.
The timing matters too. Muse launched recently and shot straight to the top of the App Store. Any credible privacy concern at that stage can shape how millions of new users think about the product going forward.
Meta's three-step defense
Meta's response came fast and from two directions. Andy Stone, Meta's VP of Communications, posted on X that the Messages integration is entirely opt-in and that Muse cannot read messages unless a user manually enables both Full Disk Access and the Messages connector. That's not a buried setting buried in a submenu. It requires deliberate action.
David Singleton, an executive at Meta Superintelligence Labs, went further on Threads. He outlined three separate permission steps spanning both application-level and macOS system-level settings, including a manual confirmation inside macOS Settings and a full app restart. His argument: even a bug couldn't bypass that chain, because the permissions are enforced at the operating system level, not just inside Muse itself.
Here's the thing, though. Aten says Full Disk Access was off. Singleton's response to that was to suggest Muse gave an incorrect explanation of its own behavior, essentially saying the AI was confused about what it was actually doing. That's a somewhat uncomfortable defense for an AI product, even if it's technically plausible.
Not the only incident raising eyebrows
Aten's piece isn't the only data point in circulation. YouTuber Matt Robb separately reported that Muse shared his home address during a Marketplace transaction. Meta's explanation there was that Robb had granted the relevant permission, which he later acknowledged. That case resolved more cleanly, but it adds to a pattern of users being surprised by what Muse can access.
The backdrop here is significant. Meta carries real baggage on data privacy, including a New Mexico jury finding tied to the Cambridge Analytica fallout. That history means any ambiguity around consent gets scrutinized harder than it would for a company without that track record. Users aren't wrong to apply extra skepticism.
What this means for anyone using Muse
The key here is understanding that "opt-in" and "clearly communicated" aren't always the same thing. Meta's permission architecture may be technically sound, but if users are regularly surprised by what Muse can see, the onboarding flow isn't doing its job.
Singleton pointing to a bug bounty program as evidence of security rigor is a reasonable signal, but it also implicitly acknowledges that bugs are possible. Aten's experience, whatever its technical cause, landed as a privacy violation from the user's perspective. That gap between technical reality and user perception is where trust erodes.
For anyone spending time across gaming-adjacent platforms and apps right now, questions like these are worth tracking. AI agents with broad device access are becoming standard features, not edge cases. Check out our gaming guides for more coverage on the tools and platforms shaping the space.
Muse's App Store position suggests most users aren't deterred yet. But the conversation between Meta and its critics is far from settled, and how the company handles the next incident will say more than any statement about permission architecture. If you're exploring new games in the meantime, the MIO: Memories in Orbit Part 2 walkthrough and the Mewgenics Wall of Flesh guide are worth a look while this story develops.
