Openai has added to Chatgpt an agent who is able to work for you and take care of complex tasks using his own computer.
This new Chatgpt agentwhich is thinking to increase user productivity and help you in all kinds of daily tasks, is beginning to deploy for users of the Pro, Plus and Team plans in Chatgpt, in countries like the United Kingdom, and will gradually reach the rest of the markets. The exact dates in which it will arrive in Spain is not yet known, although the company has indicated that it will report them in the next few days.
As Openai announced, the new Chatgpt agent performs complete tasks combining the previous tools as Operators and deep research tools with the chatgpt conversational model, which allows it to navigate the web, reason and execute tasks using its own virtual browser.
Thus, the new chatgpt agent adapts to the user's real needs and time, helping him in his daily tasks. For example, you can perform day -to -day basis such as filling online forms, making the purchase, creating documents, presentations and spreadsheets, and carrying out thorough investigations on several web pages.
To use it, it is necessary to choose the "agent mode" in the chatgpt tool selector. Done this, you just have to provide you with the necessary orders (prompts) for you to attend your work and daily chores.
For example, you could ask you to collect global data on the cost of mobile phone taifas and establish average prices for each country and that perform a ranking and present it on a slide that shows different concepts (prices, data offered, calls included, etc ...) highlighting the first 10 countries of the classification. All in the same order.
In addition, the chatgpt agent, to guarantee safety and privacy, requests explicit permission before performing any sensitive action, such as providing personal data, and also performs high -risk tasks, such as financial transactions or giving legal advice.
In the same way, it is trained to recognize and reject malicious instructions and alert the user to any ambiguity or risk that may occur. For critical actions - as the sending of an email - approval must always be given and does not act autonomously, according to OpenAI.
Leave a Reply