October 8 News.GooglePreviewing a whole new one Gemini ARTIFICIAL INTELLIGENCE MODEL, WHICH AIMS TO ENABLE AI TO BROWSE AND INTERACT WITH THE NETWORK THROUGH A BROWSER AgentIt can perform operations in user interfaces originally designed for people rather than robots. This model, Gemini 2.5 Computer Use, uses " visual understanding and reasoning " to analyse user requests and perform corresponding tasks, such as filling in and submitting forms。
The model can be used for user interface testing or for operating systems that are only for human users, do not provide API or other direct interfaces. Previously, such models had been applied to smart body functions in Google AI Mode, as well as to the research prototype project "Mariner", which used AI smarts to perform autonomous tasks in browsers, such as automatically adding goods to shopping carts on the basis of food lists。
Google's launch coincides with the day after OpenAI announced the launch of ChatGPT's new application on its annual Dev Day. OpenAI is continuously focusing on its "ChatGPT Smart" function, which allows users to perform complex tasks. Meanwhile, Anthropic has released a "computer use" version of its Claude AI model last year。