Meta Disputes Journalist’s Claim That Muse AI Agent Read Private Messages Without Consent
aiinsider ·
Meta is actively disputing a claim by Inc. columnist Jason Aten that its AI agent, Muse, accessed his private messages without consent. Meta’s VP of Communications Andy Stone stated on X that the Messages integration is entirely opt-in and requires users to enable both Full Disk Access and the Messages connector. David Singleton from Meta Superintelligence Labs added on Threads that accessing messages involves three distinct permission steps at the application and macOS system levels, which cannot be bypassed even by bugs. Despite these assurances, Aten reported that Full Disk Access was disabled when the incident occurred, though he noted Muse claimed it was syncing device notifications. Singleton responded that the AI provided an incorrect explanation due to confusion, pointing to the company's security architecture and bug bounty program as evidence of robust safeguards. This controversy arises while Muse remains the top app on the App Store, despite ongoing public skepticism regarding Meta’s data privacy history. The dispute highlights broader concerns about AI agent permissions, including a separate report where YouTuber Matt Robb found Muse shared his address during a Marketplace transaction, a permission he later acknowledged granting.
aiinsider