You Can Now Let Claude Take Over Your Computer. Here’s Why That’s Both Exciting and Terrifying.

Anthropic just handed Claude the keys to your computer. The chatbot can now take control autonomously, moving your cursor, typing on your keyboard, opening files and navigating programs like a human would.

The new feature is Anthropic’s answer to AI agents like OpenClaw. Claude looks for the right tools to complete tasks via connectors with apps like Google Calendar or Slack. If a connector isn’t available, it can manually perform the task by clicking and scrolling. The AI will always ask for permission first, and you can stop it at any time.

But the risks are real. Anthropic warns the feature may contain errors, and security experts say AI agents can be hijacked to access your personal data or execute unwanted actions. The company strongly advises against using apps with sensitive data — some are disabled by default for safety. It’s available for Claude Pro and Max subscribers on MacOS only.