A big change has started in the world of tech. Till now, we used to ask questions to AI and get answers in return. But, now AI is not limited to just giving answers. Anthropic, a well-known American AI company, has introduced a groundbreaking feature in its Claude AI. With the help of this feature, AI can directly take control of your computer and perform many tasks automatically. What’s more, in the last six months, several companies have launched “agentic” web browsers. These browsers take control of the browsing interface to perform various tasks on your behalf. Claude’s new AI tool is a step forward in this regard; It not only controls the browser to function, but takes complete control of the entire computer system. In this new era, Artificial Intelligence will no longer just give instructions; It will actually move the mouse, click, and do all the work for you. Testing of this feature was going on for some time, but now it has been officially launched. Although it may take some time to reach the general public, demos of its capabilities have already been shown.
What is this new “Computer Use” feature?
Anthropic has introduced a feature in Claude called “Computer Use”. Simply put, this AI can look at your computer screen and interact with it just like you do. Just like you use a mouse to click a button, Claude can do the same thing. It can type, open applications, and save files using the keyboard. In fact, it works not just within a particular software application, but across the entire computer system.
How does “Computer Use” work?
The real strength of this feature lies in Claude’s ability to understand the screen interface. It recognizes text, buttons, and applications that appear on the screen. Then, it works based on specific instructions you give. For example, suppose you instruct it to extract data from a particular website. Claude will automatically open a web browser, go to that website, click on the right things, and pull out the information you need. Similarly, if you instruct it to create and save a file, it will open an application like Notepad or a document editor, type the contents itself, and then save the file. In other words, tasks you would normally do in 10 to 15 steps can now be done automatically by AI, saving you time. You can automate most of your tasks.
Is this the beginning of a new AI
Generally, ordinary people use AI mainly to ask questions. For example, if someone asks, “How do I write an email?”, the AI provides the necessary text. But, the new Claude can now automatically open your email client, compose messages, and even send them. This means that it will do the entire process automatically. This is the biggest difference: AI is no longer just an advisory tool, but is becoming an active, functional system. Recently, Perplexity has also introduced a feature called “Computer”, which provides largely similar capabilities. This feature has tremendous potential to make everyday tasks easier. Office tasks—like data entry, creating reports, or finding information—can all be automated by AI. If you need to collect data from multiple websites, AI can speed up the process significantly. Repeated tasks—like filling out forms or creating files—become much easier for AI.
What are the risks?
Wherever there is advanced technology, there are also risks. If an AI can control your computer, it also gets access to your data. This feature requires allowing the AI to take complete control—that is, “take over”—of your computer system. In effect, you are giving power to an AI without any restrictions. As a result, the possibility of loss also increases. If incorrect instructions are given, or if something goes wrong with the system, the AI may do the wrong thing. In some cases, it has been observed that the AI misunderstands the context or inadvertently makes a wrong click. In other words, it is not completely infallible. Currently, this feature is still in its early stages and is not completely secure yet. Right now it is being tested with only a few users. The company is continuously improving the system to reduce mistakes and increase its reliability. This special feature of Claude is not just a minor update; This is the beginning of a new journey—one in which AI will no longer just work *with* you, but work *for* you. Prior to this, several other companies had already introduced their own automated, agent-based web browsers.