Meta has strongly refuted claims that its Muse AI agent read a user's private messages without permission, following a report published by Inc. columnist Jason Aten. Meta Vice President of Communications Andy Stone and Meta Superintelligence Labs executive David Singleton both pushed back against the allegations. Stone stated that the Messages integration in the Mac app is entirely opt-in, requiring users to explicitly enable both Full Disk Access and the Messages connector.
According to Meta executives, the software features multiple security barriers that prevent unauthorized data access. Singleton explained that granting permissions involves three separate steps of application-level permissions and built-in macOS system-level protections that cannot be bypassed even if a bug were present. Aten claimed that Full Disk Access was disabled when the incident occurred, and that the AI assistant explained it was syncing device notifications. Singleton countered that the AI was simply confused and provided an incorrect explanation of the event.
Despite the technical denials from Meta leadership, user trust remains fragile due to the company's historical controversies regarding consumer data mishandling, including recent legal setbacks. While the Muse application currently holds the top position on the App Store, maintaining consumer confidence will be critical for Meta's success in the artificial intelligence market. As additional user complaints surface regarding unexpected AI behavior, the company's ability to address software transparency and accountability will determine its long-term reputation.
Why it matters: User trust and data privacy permissions remain critical hurdles for tech giants deploying consumer-facing artificial intelligence applications.