The time has come: AI controls your Mac as if it were a human. The first tool that makes this possible is now available and more options will follow soon. In this article we explain what it is and what it can do for you.
AI controls your Mac: hands free
AI tools are taking over more and more tasks and the next step is that AI will control your entire Mac. AI company Anthropic now has the scoop on this with ‘computer use’. His AI model Claude can view a computer screen, move the mouse, click and press keys. Exactly the way a person uses a computer.
Although Anthropic is the first, the other major AI companies are busy releasing similar tools. OpenAI is finalizing Operator, which controls computers to execute commands. Google also appears to be at an advanced stage with its version of this technology, but may focus on Android smartphones first.
Why would you want that?
These tools are examples of AI Agentswho autonomously carry out assignments on your behalf. While any AI tool can perform a command, a AI Agent independently carry out various subsequent assignments to achieve a goal. This technology allows AI to use any apps on your computer to achieve that goal.
The possibilities are, in theory, endless. Every time you can’t figure something out, or don’t feel like it, ask the AI. Are you working with Excel and can’t figure it out? Then you could ask an AI to fill in the sheet for you in a specific way. Or you ask the AI to open a stack of images on your computer in a photo editing app, perform a series of standard actions there and then save the images in a separate folder. In theory, you can even ask an AI to regularly post to your various social media.
The technology can of course also mean a lot for the accessibility of modern technology, for people who cannot control a mouse or keyboard.
How do I get started?
Right now you can only computer use from Anthropic and that is still too complicated for most people. It’s not as easy as downloading an app and enabling it. You can here find information about how to install and use computer use. This is still beta software, which makes it slow and error-prone.
But once OpenAI and Google join the fray, you can expect these tools to become increasingly simple and accessible.
Are there also risks?
Be careful, because using such tools involves many risks. You basically give access to your personal computer with all your personal information, which is probably also connected to the internet. Although the AI Agent can theoretically accomplish many tasks with these resources, it can also do a lot of harm.
We all know how illogical an AI can sometimes behave. Think of hallucinating information to arrive at a satisfactory answer. In the case of a response to a search query, this is relatively harmless, but an irrational AI going wild on your Mac is scarier. What if you’re not looking and the AI deletes your photos, changes your email password, or maybe even does something illegal?
This is the future
Anthropic also admits that its AI sometimes takes strange actions to find a solution, but as mentioned, this is a beta version. Recent years have proven that the development of this technology can suddenly go very quickly. While it may seem strange now to give an AI access to your personal computer, in a year’s time it may well be the most normal thing in the world.
Think about which company you trust with this, or whether you wait for Apple Intelligence to support it. In any case, the first version of Apple Intelligence will come to your iPhone in April 2025. Even your car has to deal with it.
Download the iPhoned app
