Meta is challenging allegations that its Muse AI agent accessed private user messages without authorization. The controversy began after Inc. columnist Jason Aten reported that the tool read his messages, despite his claim that Full Disk Access was disabled.
Meta VP of Communications Andy Stone stated that the Muse app for Mac requires explicit user consent. According to Stone, users must manually enable both Full Disk Access and the Messages connector for the app to function as described.
David Singleton from Meta Superintelligence Labs added that the macOS architecture requires three distinct levels of permission. He noted that these system-level protections cannot be bypassed, even if the application contains a bug.
Singleton also addressed the AI's claim that it was syncing device notifications, stating that the model provided an incorrect explanation. The company maintains that the reported unauthorized access is technically impossible under current security protocols.
Separately, Meta is investigating a report from YouTuber Matt Robb, who alleged that Muse mishandled a Facebook Marketplace transaction. In that instance, the user's address was shared, leading to an unexpected visitor at his home.
