Microsoft has introduced new agentic capabilities for Windows 11 aimed at creating an AI-native operating system. A newly released support document highlights the evolution of Windows towards a “smart” OS, implementing features like the experimental agent workspace, currently available to Windows Insiders. This feature allows users to grant AI agents access to apps and files, operating in a secure, separate Windows session, thereby enhancing productivity while maintaining user control. Key security pillars include non-repudiation, confidentiality, and user authorization, emphasizing transparency and continuous security commitment. Microsoft also outlines strict guidelines for third-party developers to ensure compliance with safety protocols when integrating agentic capabilities into their apps. With Copilot Actions set to be the first AI application utilizing these features, Microsoft is determined to uphold security and trust in this innovative venture. As development progresses, the focus remains on delivering secure, efficient, and user-controlled AI experiences within Windows 11.
Source link
