Anthropic has introduced a new capability for its Claude Cowork and Claude Code assistants: autonomous computer control. The AI can now use your Mac while you’re away, opening applications, browsing the web, and accessing development tools to complete tasks.
How It Works
The feature, currently in research preview for Claude Pro and Max subscribers on macOS, requires no setup. If the AI needs a specific tool, it will click and navigate on your screen to obtain it. Anthropic highlights its compatibility with Dispatch, enabling users to assign tasks via their phones.
Safety Measures & Limitations
Unlike unconstrained automation tools like OpenClaw, Claude’s version includes safeguards. It prioritizes using existing connections (Slack, Google Calendar) and requests permission before accessing new applications. However, Anthropic acknowledges that AI can err, and evolving threats require vigilance.
The Bigger Picture
This development accelerates the trend toward AI-driven automation. While current restrictions mitigate immediate risks, the ability for AI to independently operate a computer raises questions about trust, security, and the future of work.
The key takeaway: AI’s increasing autonomy is reshaping how humans and machines interact. Proceed with caution, and ensure robust security measures are in place.
