APPLE says it will tighten macOS privacy settings to curb misuse of Full Disk Access (FDA) permissions by third‑party AI agents, in a move following the Muse incident and a broader debate about how AI assistants handle private data. The change comes after Meta’s Muse AI reportedly accessed Apple Messages, prompting questions about whether FDA alone is enough to prevent edge cases where sensitive data could be exposed, even when users have granted permissions.
The core debate centres on how Muse could read messages. Meta contends that, for Muse to access Messages, a user must enable both macOS Full Disk Access and a Messages connector in the Muse app. Critics, including security researcher Patrick Wardle, argue that FDA effectively makes any non-root data readable, raising concerns about reading messages, browsing history and other data.
Apple’s official statement emphasised that some developers misuse FDA in ways that can jeopardise user and contact privacy, and warned that as AI agents become more capable, the associated risks grow unless users are clearly informed before granting access.
While Apple did not name Muse or Meta specifically, the timing follows Wardle’s disclosure of a Muse configuration that could grant broad control to the AI assistant and recent reports that Amazon blocked Muse from its platform. The article notes there are no known public reports of other apps abusing FDA for reading messages, but the incidents have intensified scrutiny of how AI agents are granted system-wide permissions and the safeguards surrounding them.