OpenAI Operator Will Do Your Clicking and Typing for You
Openai’s new AI Tools, operators are designed to make your mundane tasks easier, including making dinner reservations, ordering groceries and filling out forms.
In the demo video, Openai Highlights When using a special browser, how AI agents interact with web pages by typing, clicking, and scrolling. You just describe the task you want to accomplish, and the agent can handle multiple requests at the same time – for example, shopping on Etsy when booking a dinner.
According to OpenAI, it can be “viewed” through screenshots and “interact” in a way that the mouse and keyboard allow in the browser. Operators who Openai describes it as “one of our first agents” are available in a limited preview.
Together with competitors like Google and Google Human Already provided Similar AI agentsOpenai is working hard to close the gap. This is also part of the bigger effort to make its generative AI more useful by automating more aspects of everyday life, and it is possible to get closer and closer to fulfilling the promise that we will forever change the way we interact with technology.
“The tools that are able to use the same interface and tools to interact with humans every day expand the practicality of AI, helping people save time on daily tasks while opening up new opportunities for businesses to engage,” the company said in a blog post.
The tool is powered by a new model called Computer Usage Agent, which combines GPT-4OVisual functions of advanced reasoning through enhanced learning. It is trained to interact with a graphical user interface, including buttons, menus, and text fields that people see on the screen.
The company said that if there is a problem, the operator can use its reasoning function to self-correct or return to control the user. It is also trained to require users to take over tasks that require certain inputs, such as login credentials or payment details.
The tool can now pay to professional users in the United States operator.chatgpt.com.
