Meta Denies Claims That Muse AI Accessed User Messages Without Consent

30 Sept 2026, 9:54 pm · 1d ago · 1 min read · TechCrunch AI

Meta has officially denied allegations that its Muse AI desktop assistant accessed a journalist's private messages without obtaining proper user authorization. The controversy arose after a reporter reported that the assistant interacted with his personal message history while the required operating system permissions on macOS were toggled off. Meta clarified that the Muse agent cannot read private communications without explicit user consent and stated that the observed behavior was technically impossible under its current privacy architecture. The incident highlights mounting public scrutiny and privacy concerns surrounding desktop-integrated AI agents capable of monitoring system-level activities and personal communications.