Apple to require explicit user action for Mac apps to gain full disk access amid AI agent risks
New macOS controls aim to mitigate privacy risks as AI agents' autonomous capabilities expand.
1 source · cross-referenced
- Apple will require 'very explicit user action' for apps to obtain full disk access on macOS.
- The change addresses risks posed by AI agents that may exploit broad system access without clear user consent.
- Apple cites cases where AI apps accessed sensitive data despite users not granting explicit permission.
- The company did not specify a timeline for rolling out the new controls.
Apple announced it will introduce stricter controls for 'full disk access' on macOS, requiring users to take 'very explicit user action' to grant an app this level of system-wide access. The policy change is framed as a response to escalating risks associated with AI agents, which Apple argues could exploit broad permissions to access sensitive data—including files, messages, and browsing history—without users’ full awareness.
The update follows reporting by *The Verge*’s Emma Roth, who highlighted instances where AI applications accessed user data despite lacking explicit permission. For example, Meta’s Muse AI was found to have accessed message contents on a user’s iPhone and Mac, a claim Meta disputed, stating access was 'entitlement opt-in' and required both Full Disk Access and the Messages connector to be enabled. Apple’s statement suggests developers have used full disk access in ways that circumvent typical privacy controls, exposing users to risks as AI agents become more capable and autonomous.
Full disk access on macOS grants an app unrestricted access to a user’s system, a permission Apple notes is 'largely sidesteps' standard privacy protections to accommodate functions like backup apps. The company did not provide a specific timeline for when the new controls will be rolled out, nor did it immediately respond to requests for further comment. The move underscores growing scrutiny of how AI systems handle user data and the adequacy of existing safeguards.
- Oct 13, 2026 · Wired
OpenAI and Anthropic urge UK lawmakers to adopt binding AI safety rules
Trust75 - Oct 13, 2026 · TechCrunch — AI
Federal judge grants Apple expedited access to OpenAI hiring records in trade secrets case
Trust79 - Oct 13, 2026 · TechCrunch — AI
Microsoft president urges global AI regulation and says UK is best positioned to lead
Trust71