TipsMake

Claude AI can use computers like humans, assisting in tasks that replace human users.

Anthropic adds computer control capabilities to Claude, allowing the AI ​​to automatically send files, edit photos, and handle tasks on behalf of the user.

AI assistants are taking a big step forward. Anthropic has just announced a new capability for Claude, allowing the AI ​​to directly control the computer to perform tasks on behalf of the user .

 

This means you can assign Claude everyday tasks like sending files from your computer to your phone, batch editing photos, or checking email — and the AI ​​will do it for you.

Claude can use computers just like a human.

According to Anthropic, Claude can manipulate computers like a real user. This AI can open browsers, use developer tools, access open files, and interact with applications on the system.

Claude can even move the mouse, type text, and scroll pages like a normal user. Additionally, the system supports connections with services like Google Calendar or Slack to perform automated tasks.

In the absence of a connection, Claude can still manually perform tasks using the keyboard and mouse. Importantly, Claude will always request permission before taking any action. Users can monitor the process and stop the task at any time.

Claude AI can use computers like humans, assisting in tasks that replace human users. Picture 1

 

Additionally, the computer control feature works in conjunction with Dispatch, allowing users to assign tasks from their phone. For example, you can ask Claude to check emails every morning, create work summaries, or run software tests. This allows Claude to work even when you're not sitting at your computer.

Convenient, but also comes with risks.

While AI-powered computers offer many conveniences, they also raise security concerns. Experts warn that AI agents can perform large-scale actions in a short period of time, and if exploited, could pose a risk to personal data.

Anthropic stated that they have implemented numerous security measures such as prompt injection detection, vulnerability scanning, and access restrictions for sensitive applications. Some applications that handle critical data are even disabled by default to ensure security.

However, Anthropic still advises users to exercise caution as this feature is currently in the testing phase.

Currently, the new feature is only available for Claude Pro and Claude Max users and only supports macOS computers. Anthropic states that this is only a research/test version, so some complex tasks may not work perfectly.

Claude's new capabilities demonstrate that AI is rapidly evolving from chatbots to full-fledged automated assistants. Instead of simply answering questions, AI can now directly act on the computer and handle tasks in place of humans.

If this trend continues to develop, the prospect of AI becoming a true 'personal assistant' could soon become a reality.

Discover more

Claude AI Anthropic Claude AI agent Claude computer use AI automation
Kareem Winters

Share by

Kareem Winters
Update 30 March 2026