Just recently, Windows President Pavan Davuluri announced that Microsoft aims to transform Windows into an agent-based OS, and now we can see how this will be implemented. Although, for now, it’s only a preview. In the Windows 11 Insider Build 26220.7262, a new system setting called Experimental agentic features has appeared. The corresponding switch can be turned off, but it will hardly result in any changes. 
Created by Grok. With this feature enabled, the system creates a separate workspace called Agent Workspace. As Microsoft explains, this is a kind of Windows ‘sandbox’. This environment is isolated from the rest of the system and is supposed to be more efficient than using virtualization. However, at present, no functions are working within it just yet. It is known that every time such a separate workspace is created, a unique AI agent will be assigned, each having customizable access to a variety of key folders, including ‘Downloads’ and ‘Desktop’.
Microsoft refers to these agents as AI that can perform a range of functions on behalf of the user, although they will require various permissions to do so. For example, the AI might automatically process incoming emails and respond based on user settings, move files, launch certain applications, and so on.
Users on the social network X have generally reacted negatively to Microsoft’s related announcements, claiming that this turns Windows into a monstrous entity and that Microsoft can no longer be trusted.
Recent trends in the tech industry show that privacy concerns are growing as AI becomes deeply embedded within operating systems. Tech giants like Apple and Google are also exploring similar advancements in AI integration, focusing heavily on user consent and transparent data usage policies. Microsoft is likely to face increasing scrutiny from regulators and privacy advocates as it pursues its aggressive AI strategy.
