Apple launches new security controls for Mac to address AI agent risks
Apple dévoile de nouveaux contrôles de sécurité pour Mac face aux risques des agents IA
Apple announced an overhaul of Mac security settings, splitting device control and data access into two independent settings to address the privacy risks of AI agents directly operating the operating system. The two settings are "System & App Protection Override" and "Accessibility"; users select apps in a dedicated list to grant full disk access, and Apple recommends granting it only to software that genuinely needs it. The new settings will also show the scope of data an app can control, support requesting permission on every access, and display a transparency notice when one app controls another. These changes will ship with macOS 27.3 and will not prevent the installation of agents that require high privileges; backup tools and assistive technologies can still apply for entitlement exceptions. Previously, a journalist said Meta's Muse app read their private messages without authorization, and Wired also reported a vulnerability in the ChatGPT Mac app that could let hackers access sensitive data.