Meta Denies AI Agent Muse Read Private Messages Without Permission
Meta is denying claims that its new AI agent, Muse, accessed private messages without user consent. The dispute began after journalist Jason Aten reported an incident where the app appeared to read his iMessages despite not having full access permissions enabled. Meta's VP of Communications, Andy Stone, clarified on X (formerly Twitter) that accessing Messages requires users to explicitly enable both Full Disk Access and a specific connector within Muse’s settings. Despite these assurances from Meta, skepticism remains due to the company’s history with data privacy issues, including past lawsuits like those stemming from the Cambridge Analytica scandal. The controversy could impact consumer trust in Muse as it competes for dominance in the AI market. Another user also reported a separate issue involving Muse mishandling Facebook Marketplace information. Meta's technical team has provided detailed explanations of how permissions work and dismissed Aten’s account, suggesting his experience was either misunderstood or misrepresented by the AI itself.
Tickers: $META
Topics: AI consumer trust Meta Muse app privacy