This photograph shows a figurine in front of the logo of the AI assistant “Claude” built by the American security and artificial intelligence research company Anthropic during a photo shoot in Paris on February 13, 2026. (Photo by Joel Saget / AFP via Getty Images)
Joel Saget | Afp | Getty Images
Anthropic’s Claude can now use a person’s computer to perform tasks as the company seeks to create an AI agent that can rival the OpenClaw virus.
Users can now send a task to Claude from a phone, and the AI agent will then complete that task, Anthropic announced Monday.
After being prompted, Claude can open applications on your computer, navigate in a web browser and fill out spreadsheets, Anthropic said. One prompt demonstrated by Anthropic in a video posted Monday is of a user being late for a meeting. The user asks Claude to export a pitch deck as a PDF file and attach it to a meeting invitation. The video shows Claude completing the task.
Anthropic’s latest update highlights AI companies’ push to create “agents” that can autonomously perform tasks on behalf of users at any time of the day.
Agent abilities have been thrust into the spotlight this year following the release of OpenClaw, gone viral. OpenClaw is linked to AI models from OpenAI and Anthropic. A user can message OpenClaw through popular apps like WhatsApp or Telegram to complete tasks. Like Anthropic’s new feature, OpenClaw runs locally on a user’s device, giving them access to files.
Read more CNBC tech newsNvidia CEO Jensen Huang told CNBC last week that OpenClaw is “definitely the next ChatGPT” as tech companies rush to create their own competitors. The chip leader announced last week NemoClawan enterprise version of OpenClaw.
OpenAI hired Peter Steinberger last monththe creator of OpenClaw, as the company seeks to “drive the next generation of personal agents.”
BackupsAnthropic warned that computer use “is still early days compared to Claude’s ability to code or interact with text.”
“Claude can make mistakes, and while we continue to improve our protection measures, threats are constantly evolving,” Anthropic warned.
The company added that it had built the computer’s use capacity “with safeguards that minimize risk” and that Claude would always ask permission before accessing new applications.
Users can use Dispatch, a feature released last week in Claude Cowork. This allows users to have an ongoing conversation with Claude from a phone or desktop and assign tasks to the agent.
