Artificial Intelligence

Apple Announces New macOS Security Controls to Tame AI Agent Privacy Risks

· GlobeWire Daily Editorial Desk · Source: techcrunch.com

2 min read

Key takeawaysApple unveiled a set of macOS security updates aimed at giving users clearer control over AI agents that request full‑disk access. The changes arrive with macOS 27.3 and target the growing privacy concerns around locally‑run AI assistants.
Apple Announces New macOS Security Controls to Tame AI Agent Privacy Risks

Image: Sarah Perez · Photo source · Resized and letterboxed to 16:9; no AI generation

Summary

Apple unveiled a set of macOS security updates aimed at giving users clearer control over AI agents that request full‑disk access. The changes arrive with macOS 27.3 and target the growing privacy concerns around locally‑run AI assistants.

What Apple Is Changing

  • Two new permission groups – *System & App Protection Override* and *Accessibility* replace the old “Device Control” and “Data Access” settings.
  • Single‑click permission picker – Users will select an app from a consolidated list when granting any elevated rights.
  • Granular prompts – An optional mode forces apps or AI agents to ask for permission each time they need access, preventing blanket approval.
  • Transparency indicator – A visual cue appears when an app is controlling another app via the Override setting.
  • Improved visibility – Apple will show a summary of the data and system resources an app can reach when full permissions are granted.

Why the Update Matters

AI agents differ from chatbots: they run on‑device and can act across the OS, files, and other apps. Recent reports—such as a claim that Meta’s Muse read private messages and a Wired investigation of a flaw in ChatGPT’s macOS client—highlighted how broad permissions can be abused. Apple’s new controls aim to make those risks more visible and manageable without outright blocking AI functionality.

Impact on Developers and Users

  • Developers must revise onboarding flows to explain the new permission groups and handle limited‑folder access gracefully. Those that need Full Disk Access for legitimate purposes (e.g., backup tools or assistive tech) can request an “entitlement” exception.
  • Users gain more nuanced choices: they can allow a one‑time permission, limit an agent to specific folders, or deny Override entirely. The trade‑off is a potentially more tedious workflow for highly automated AI tasks.

How to Verify the Changes

The updates will be included in the macOS 27.3 release, scheduled for later this year. Apple’s official blog post (linked from the macOS security page) lists the exact settings and provides screenshots. Checking the System Settings → Privacy & Security section after updating will show the new groups and the transparency indicator.

Practical Takeaways

  • Expect a brief learning curve when configuring AI agents on macOS.
  • Organizations deploying local AI models should audit which apps request Full Disk Access and test the new permission prompts.
  • Keep macOS updated to benefit from the visibility and control features as soon as they become available.

--- *Source: TechCrunch, 16 Nov 2026*

Share this article
Read the original source: Apple reveals new Mac security controls as AI agents raise privacy risks

AI-assisted reporting; source claims are attributed and have not been independently verified.

About the author

GlobeWire Daily Editorial Desk — AI-assisted news based on attributed published sources.

More to read