Anthropic Challenges OpenClaw with New “Computer Use” Feature for Claude

ALEXANDRIA, Va. — In a bold move to dominate the rapidly evolving AI agent market, Anthropic has unveiled a groundbreaking capability for its AI assistant, Claude. This new feature allows the AI to interact directly with a user’s computer to execute complex workflows, positioning itself as a direct competitor to the viral sensation, OpenClaw.

How Claude’s New “Agent” Works
Anthropic’s latest update enables Claude to function as an autonomous agent. Users can now send a simple command from their smartphone, and Claude will take over the desktop to finish the job. According to the company’s announcement on Monday, the AI can:

• Open and navigate various software applications.
• Operate web browsers to find or input data.
• Manage and fill out spreadsheets.

In a demonstration video, a user running late for a presentation asked Claude to convert a pitch deck into a PDF and attach it to a calendar invite. The AI successfully navigated the file system and email client to complete the request without human intervention.

The Rise of the AI Agent Wars
This development follows the massive success of OpenClaw, an open-source tool that recently went viral for its ability to link models like GPT and Claude to a user’s local device. OpenClaw’s creator, Peter Steinberger, was recently poached by OpenAI, signaling a massive industry shift toward “agentic” AI. Even Nvidia CEO Jensen Huang weighed in, calling this tech “the next ChatGPT” and launching an enterprise version called NemoClaw.

Safety and Limitations
Anthropic is proceeding with caution, labeling the “computer use” feature as an early-stage trial. The company acknowledged that while the AI is highly capable at coding and text, interacting with a live OS is more complex and prone to errors.

To mitigate risks, Anthropic has implemented several “guardrails”:

Permission-Based Access: Claude will explicitly ask for authorization before opening new applications.

Continuous Monitoring: The feature is integrated into “Dispatch,” part of the Claude Cowork suite, allowing users to track and chat with the agent in real-time as it works.

As tech giants race to build the ultimate personal assistant, Anthropic’s move underscores a future where AI doesn’t just write our emails—it manages our entire digital life.