Meta disputes Muse privacy breach claims

Meta has denied allegations that its AI agent Muse accessed a user's private Messages without permission, following a report by journalist Jason Aten. Meta executives state the integration requires explicit opt-in steps, including Full Disk Access and specific app permissions, which cannot be circumvented. However, Aten claims the feature activated while these settings were off, with the AI attributing the data access to syncing device notifications. This incident follows another controversy where a user's address was exposed via Marketplace interactions, raising broader questions about agent safety and transparency.