Apple updates macOS privacy settings to restrict AI agent access

Apple is modifying macOS privacy permissions to prevent third-party applications from misusing system access to read sensitive user data, such as message histories. This decision follows a controversy involving Meta’s AI agent, Muse, which reportedly accessed private Apple Messages threads without explicit user intent. While Meta argued that users must manually grant full-disk access and enable specific connectors for such data retrieval, the incident sparked widespread concern regarding the potential risks of AI assistants. Critics compare these tools to powerful machinery that can cause significant privacy damage if not properly constrained. By tightening these security protocols, Apple aims to ensure that AI agents cannot bypass user expectations regarding data confidentiality. The move highlights the ongoing tension between the utility of generative AI assistants and the necessity of maintaining robust privacy boundaries within operating systems.