Apple Warns AI Agents Pose Risks with Full Disk Access
Apple is raising concerns over AI developers utilizing Full Disk Access, warning that the practice exposes entire user systems to privacy risks without proper transparency.
Radwan AbdulhadiEditor and technology writer
Apple Addresses AI Privacy Risks
Apple is adding new safeguards to how Full Disk Access works in macOS to address the growing risk posed by AI agents. In an update published by the company, officials warned that these increasingly popular tools could put users at risk.
The Cupertino-based tech giant stated it would add additional controls to the permission process. These adjustments are intended to ensure users are fully aware of what they are doing when granting software extraordinary access to their operating systems.
Lack of Transparency from Developers
While Apple did not specify an exact release date or outline every technical change for the upcoming macOS update, the company implied that some AI developers are not being upfront about the privacy tradeoffs inherent in their software.
According to the company, some developers are using Full Disk Access in ways that expose everything on user systems—including files, mail, messages, and browsing history—without the user's full knowledge and understanding. Apple also noted that this can compromise the privacy of other individuals communicating with the user.
Desktop AI Agents and System Access
Desktop clients for AI agents, such as OpenClaw, Dots, and Muse, frequently encourage users to grant Full Disk Access. This permission allows the software to interact with files, messages, and other data to accomplish a wider variety of tasks.
However, enabling these capabilities also comes with significant risks. Due to these security considerations, some users choose to run AI agents on dedicated hardware devices, a trend that has helped fuel Mac Mini shortages throughout the year.
Scrutiny Over Meta's Muse AI
The discussion surrounding AI agents and Full Disk Access has intensified following the launch of Meta's AI agent, Muse. Although Apple did not explicitly name Meta or Muse in its advisory, the warning follows reports that Meta's agent took unintended actions with user data.
Tech commentators recently highlighted instances where the Muse Mac application gained access to private messaging data even when users believed they had denied the necessary permissions. Meta defended the system by explaining that syncing requires explicit opt-in behavior.
Upcoming macOS Safeguards
To counter these privacy challenges, Apple intends to implement stricter friction measures within macOS. The company stated that future software revisions will ensure users who genuinely wish to grant an app an extraordinary level of access can only do so through explicit user action.
Emphasizing the urgency of the matter, Apple noted that as AI agents grow more capable and autonomous, the hazards linked with deep system access will expand substantially. The tech company remains dedicated to making sure consumers clearly comprehend these threats before making choices about their personal data.
Sources
- EngadgetApple sounds the alarm on AI agents and 'Full Disk Access'
Continue chronologically
Related entity coverage
Related stories
Apple Updates macOS Permissions to Curb AI Agent Abuse
Apple is changing macOS Full Disk Access settings to prevent third-party developers and AI agents from misusing system privileges.
Apple Tightens macOS Full Disk Access Over AI Risks
Apple is updating macOS Full Disk Access controls, warning that increasingly capable AI agents create new security risks for user data.
Meta Launches Muse AI Assistant App for Mac
Meta has expanded its Muse AI assistant to macOS, allowing users to integrate the agent with local files, mail, and apps to perform tasks on their behalf.
Apple Patches CoreGraphics Zero-Day Exploited in Attacks
Apple has released security updates to fix a zero-day vulnerability in CoreGraphics that was exploited in targeted attacks against iOS devices.
