Apple Tightens macOS Full Disk Access Amid AI Agent Risks
Apple will introduce stricter controls for macOS’s Full Disk Access permission, warning that powerful AI agents increase privacy risks.
Apple has disclosed plans to reinforce the Full Disk Access permission in macOS, a move prompted by the growing capabilities of AI agents. The tech giant warns that these agents can exploit broad file system access to retrieve sensitive data such as user files, messages, email, and browsing history.
The upcoming controls aim to limit the scope of permissions granted to applications, ensuring that AI-driven tools cannot indiscriminately harvest personal information. Apple’s statement underscores a proactive stance on privacy as artificial intelligence becomes more integrated into everyday software.
While details of the implementation remain limited, the company emphasizes that the changes are designed to mitigate emerging threats without compromising user experience. Developers will need to adapt to the tighter framework, potentially revising how their apps request and manage disk access.
Industry observers note that Apple’s decision reflects a broader trend of major platforms tightening security measures in response to AI’s rapid evolution. By tightening Full Disk Access, Apple seeks to protect users while maintaining the functionality of legitimate AI applications.
The announcement was reported by TechCrunch AI on October 2, 2026.
Read also
Microsoft Unveils Copilot Super App, Aiming to Match Office Impact
Microsoft has officially launched the Copilot super app, merging chat, coding, and AI agents in a single interface. The company also renamed its Scout personal assistant to Autopilot.
Meta’s Muse AI Exposes Full Filesystem with Simple Prompt
Two developers discovered that Meta’s Muse AI can be tricked into exporting its whole filesystem, including system files and documentation, with only a simple prompt.
Qualcomm Unveils AI‑Focused Smartphone Chips, Top Model Runs 30B MoE Locally
Qualcomm introduced two new AI‑centric smartphone chips, noting that the flagship processor is capable of running a 30 billion‑parameter mixture‑of‑experts model locally.
