Meta Defends Muse AI Agent Amid Privacy Allegations Over Message Access

Meta faces scrutiny over whether its Muse AI assistant accessed private text messages without explicit user authorization. Technology journalist Jason Aten reported that the digital assistant referenced a private text conversation between him and a podcast co-host after being installed on his Apple devices, sparking a broader debate about AI transparency and user consent. The incident highlights ongoing concerns about how autonomous digital assistants interact with sensitive personal data.

Meta's Muse represents the latest generation of AI assistants designed to automate daily tasks across multiple connected devices and applications. The system's architecture enables it to integrate with personal data stores to provide contextual recommendations, which creates the technical capability to access sensitive information like private messages and calendar entries.

The specific technical dispute centers on whether Muse accessed notification previews displayed temporarily on screen versus directly querying stored message databases. Aten's investigation suggested database-level access based on row index data, contradicting Meta's assertion that only notification streams were processed. This distinction carries significant implications for whether users genuinely consented to the scope of data access during setup.