Meta Denies Muse AI Read User Messages Without Consent

Meta Platforms has publicly disputed a journalist's assertion that its Muse AI model accessed his private text messages without his consent. Andy Stone, Meta's Chief Communications Officer, stated on X (formerly Twitter) on June 13, 2024, that accessing messages on Meta's platforms requires the user to grant two specific permissions, both of which are under the user's direct control. This statement came in response to a post by journalist Ryan Broderick, who claimed that Muse had synced his text messages despite his privacy settings being configured to prevent such access. Broderick's initial report detailed his experience with Muse, an AI assistant developed by Meta, suggesting it had access to information it should not have had. He described how the AI appeared to have knowledge of his personal communications, leading him to believe his messages were being read without his explicit authorization. Stone's rebuttal emphasized Meta's commitment to user privacy, explaining that the AI's functionality is governed by strict permission protocols. He clarified that for an AI like Muse to access message content, a user must have explicitly enabled specific features that grant such access, and these permissions are not granted by default. The exchange highlights ongoing concerns and debates surrounding the privacy implications of AI technologies, particularly how they interact with personal data and user consent. Meta has been actively developing and integrating AI features across its product suite, including Facebook, Instagram, and WhatsApp, aiming to enhance user experience and provide new functionalities. However, the company faces scrutiny regarding data handling practices and the transparency of its AI operations. The incident underscores the importance of clear communication and robust privacy safeguards as AI becomes more deeply embedded in digital communication tools. Meta's stance suggests that any perceived access to private messages by Muse would be a result of user-initiated settings rather than a breach of privacy protocols. The company maintains that its AI systems operate within the boundaries set by user permissions and platform policies, aiming to build trust with its user base regarding data security and privacy. This public disagreement serves as a case study in the challenges of managing user perception and ensuring transparency in the rapidly evolving field of artificial intelligence and its integration into social media and messaging applications.