Summary
Anthropic has introduced a new feature that allows its AI tools, Claude Code and Claude Cowork, to take direct control of a user's computer desktop. This update enables the AI to move the cursor, click buttons, and type just like a human would. The goal is to help users finish complex tasks by letting the AI navigate through different apps and files on its own. While this technology is still in a testing phase, it marks a major step toward AI becoming a more active assistant in daily work.
Main Impact
The biggest change with this update is that the AI is no longer stuck inside a simple chat window. By gaining the ability to "see" and interact with a computer screen, Claude can now handle jobs that involve multiple programs. For example, it can open a web browser to find information, copy that data into a document, and then use a coding tool to update a file. This reduces the need for users to manually move data between different apps, making the AI a much more powerful tool for developers and office workers alike.
Key Details
What Happened
Anthropic announced that its specialized tools, Claude Code and the more user-friendly Claude Cowork, have been updated with "computer use" capabilities. When the AI needs to finish a task, it can now ask for permission to navigate the user's screen. It does this by looking at what is visible on the monitor and deciding where to click or what to type. This feature is designed to work when there is no direct link between the AI and a specific app. Instead of waiting for a special update for every piece of software, the AI simply uses the computer the same way a person does.
The company also mentioned a tool called Dispatch. This allows a person to send tasks to their computer from a different location. As long as the main computer is turned on, a user can tell Claude to start a task remotely, and the AI will begin working on the desktop as requested.
Important Numbers and Facts
Currently, this feature is not available to everyone. It is being released as a "research preview," which means it is still being tested and improved. Here are the specific requirements and facts regarding the launch:
- The feature is only available for users on MacOS at this time.
- Users must have a Claude Pro or Claude Max subscription to access these tools.
- Anthropic warns that using the computer directly is slower than using direct app connections.
- The AI may make mistakes and might need a second try to finish difficult jobs.
- Safety measures require the AI to ask for permission before it starts clicking and scrolling on the machine.
Background and Context
For a long time, AI has been used mostly to write text or answer questions. However, the tech industry is now moving toward "AI agents." These are programs that can actually do work instead of just talking about it. Other big tech companies are also working on similar tools that can control a mouse and keyboard. Anthropic is trying to stay ahead by giving Claude the ability to handle the messy reality of a standard computer desktop, where things are not always organized in a way that software can easily understand.
Before this update, Claude mostly relied on "Connectors." These are direct digital bridges to specific apps like Google Drive or Slack. Connectors are very fast and safe because the AI doesn't have to "look" at anything; it just sends data back and forth. However, many apps do not have these bridges. By adding the ability to use the screen directly, Anthropic ensures that Claude can work with almost any piece of software ever made.
Public or Industry Reaction
The reaction to this news has been a mix of excitement and caution. Experts note that giving an AI control over a computer is a big responsibility. Anthropic has been very open about the current limits of the system. They have stated clearly that this method is more "error-prone" than their other tools. By calling it a research preview, they are telling users to expect some bugs. This honest approach is seen as a way to manage expectations while still showing off what the future of work might look like.
What This Means Going Forward
In the future, we can expect AI to become even more integrated into our computers. As the software gets faster and more accurate, the "slow" feeling Anthropic mentioned will likely disappear. However, this also brings up important questions about security. If an AI can click anything on a screen, companies will need to make sure it cannot be tricked into doing something harmful. For now, the requirement for user permission is the main safety net. As these tools move out of the testing phase, we will likely see them arrive on Windows and other operating systems as well.
Final Take
Anthropic is pushing the boundaries of what a digital assistant can do. By allowing Claude to step out of the chat box and onto the desktop, they are making it possible for AI to handle real-world tasks that were previously too complex. While it is still early days and the system has some flaws, the ability for an AI to navigate a computer screen marks a turning point in how we use technology to get things done.
Frequently Asked Questions
Can Claude control my computer without me knowing?
No. The system is designed to ask for permission before it begins to scroll, click, or explore your desktop. You must grant access for the AI to start working on your machine.
Is this feature available on Windows?
At the moment, the computer use feature is only available for users on MacOS. Anthropic has not yet announced a specific date for when it will be available for Windows users.
Why is using the screen slower than using a direct app link?
When the AI uses the screen, it has to take screenshots, analyze what it sees, and then decide where to move the mouse. This takes more time and computer power than a direct data connection between two programs.