Technology news and discussion on TalkFuse.
Apple adds granular Mac permissions for AI agents in macOS 27.3
Apple is changing Mac security controls as AI agents increasingly ask for broad access to files, apps and system functions. The new settings are intended to make the risks clearer and give users more control, while still allowing them to run these tools.
Coming with macOS 27.3, the changes split existing controls, simplify how users grant permissions and add guidance about limiting access to specific apps and folders. Users will be able to require permission each time an agent needs access, and see an indicator when an app controls another app. Apple cited concerns following reports of an AI app accessing private messages and a flaw that could expose sensitive data; developers may request exceptions for backup and assistive technology apps.
In short
- Apple is updating Mac permissions as AI agents seek broad access.
- macOS 27.3 adds more granular controls and clearer warnings.
- Users can require approval each time an app needs access.
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
More granular permissions could help users understand what an AI agent can access and limit the damage if a tool behaves unexpectedly or has a flaw. Requiring approval for each access and showing when one app controls another may give people meaningful oversight while still letting them use these tools.
The case against
Frequent prompts and tighter access limits could make AI agents less useful, especially when they need to work across files and apps without constant interruption. The changes may also add complexity for users and developers, while exceptions for some apps could make the rules harder to apply consistently.
Explained simply
AI agents are software tools that can carry out tasks on a Mac, such as working with files or using other apps. To do this, they may need access to personal information and parts of the system.
Mac security settings govern what apps can access and whether they can control other apps. The debate around these tools concerns how to let them work while limiting the chance that private data is exposed or changed without the user’s knowledge.
Apple makes the Mac’s operating system, macOS, and sets the rules developers must follow when building apps for it. Changes to those rules can affect both people using AI tools and developers making them, including providers of backup and accessibility software.
Read the full article at the source →
Originally published by TechCrunch as “Apple reveals new Mac security controls as AI agents raise privacy risks”.