Anthropic has introduced a new feature for its Claude AI assistant that allows it to control a user’s computer with precision. The update, currently in a research preview stage, enables Claude to open applications, browse the internet, and execute tasks across the desktop environment.
Claude’s functionality extends to clicking, typing, and navigating through various applications, mirroring human user interactions. However, a crucial aspect of this advancement is Claude’s adherence to seeking permission before executing any action, ensuring user consent for each task. Notably, certain sensitive categories such as financial and trading platforms are automatically restricted by default to safeguard user data and privacy.
Users retain the ability to manage Claude’s memory, enabling them to view, edit, or delete stored information as needed, offering a level of control and transparency over the AI’s activities on their computer.
Source: Tech-Economic Times