Thanks to their ability to answer our questions, generative artificial intelligence models are already very useful. However, a new generation of AI is brewing. And it not only responds to prompts by generating text and images, but can also carry out concrete actions to make our lives easier. The media calls this type of artificial intelligence “AI agents”.
And while waiting for the technology to become widespread, the startup Hume AI has just published a demonstration of the capabilities of this technology. Hume AI has developed artificial intelligence for voice interactions with emotional intelligence. “We want AI to understand what frustrates and upsets you, by understanding your voice and not just what you say. It can learn from these signals to better understand your personal preferences.”explains Alan Cowen, the founder of this startup.
And to go even further, Hume AI collaborated with Anthropic, one of the most promising AI labs, to combine its technology with Anthropic’s Computer Use functionality. Launched in beta (for developers) in October, this futuristic functionality allows artificial intelligence to take control of a computer, to carry out actions for the user. In its announcement, Anthropic announced that “(…) developers can teach Claude (editor’s note, Anthropic’s AI model) to use computers in the same way as others: by looking at a screen, moving a cursor, clicking buttons and while typing text”.
Chat with your PC instead of using a keyboard and mouse
It is not certain that we could completely get rid of keyboards and mice. But in any case, it will be possible to entrust certain tasks to AI instead of manipulating these devices. In the video published by Hume AI, the tool it developed using Anthropic technologies takes control of the Mozilla Firefox browser. Using an extremely smooth and natural voice interface, the user can start a game of chess and play without touching their mouse, just by chatting with the AI.
And this is still just a glimpse of what is becoming possible with AI agents. Anthropic, during the presentation of Computer Use, indicated that other companies, such as Canva or The Browser Company (the developer of the Arc browser) are already exploring this technology.
Competing technologies under development
Currently, it is rumored that Google is working on an AI called “Jarvis” which could serve as an agent on Google Chrome, to carry out tasks on the web for the user. It is also rumored that OpenAI, the creator of ChatGPT, is also working on an AI agent which could arrive as early as next year. In other words, we could soon be entering a new era of automation.
- Thanks to its ability to respond to our prompts, generative artificial intelligence is already very useful
- However, a new generation of AI will also make it possible to automate certain tasks on our PCs
- And recently, startup Hume AI released a demo of this technology leveraging Anthropic’s Computer Use feature
- Launched in beta in October, this feature can use PCs like a human would
- Other companies, like Canva or The Browser Company (the developer of the Arc browser) are already exploring this technology
- Rumors suggest Google and OpenAI are working on similar technologies
