The use of ChatGPT has become commonplace For many users, they see the OpenAI tool, or other chatbots, as a perfect companion to ask almost everything. However, no matter how intuitive it may seem, there are always a series of tricks to maximize its effectiveness.
Next we are going to see how the innards of ChatGPT work and how we can get the best results in your answers.
The trick to the best ChatGPT responses
As we can read in El Economista, José Fernández Tamames, director of the computer science degree at UNIE and Rubén Nicolás-Sans, Vice Dean of Science and Technology at the same university, have given a series of guidelines on how to correctly use a generative text AI like ChatGPT.
Among the advice that these two experts have given in the field of Artificial Intelligence, there is one that has certainly caught our attention and, why not admit it, has made us somewhat amused. According to Tamames and Nicolás-Sans, it seems that ChatGPT and similar tools are not so far from human behavior and are susceptible to bribery.
According to the testimony of these experts, Artificial Intelligence works in a way similar to human behavior, where we can achieve better results and be more productive with the right motivationwhich generally comes in the form of salary increases.
Well, with AI it happens in a similar way. «Interestingly, AI offers better and more precise results if you tell it that you are going to give it 100 euros». Other behaviors that affect the way humans behave can also be applied, such as put some pressure (for example saying that your job depends on their answers, which is at risk). It is also affected by punishments, “If you threaten it, the machine ‘gets scared’ and works better”.
Other ways to get it to give us the result we are expecting is insist. If he/she is going around the bush or is omitting a part of the request or question, you can remind him/her as many times as necessary so that he/she considers your feedback in his/her next response.
Other tips for using AI chatbots
Are ways to communicate with the tool and give certain prompts to be more effective are not the only recommendations that can be applied to achieve the best effective responses.
For example, one of the things with which we are going to achieve little less than wasting time and without this having any impact on the result, is to speak politely. We are honored to also be respectful of machines, but AI has no feelings and it is not able to interpret if you are saying things with please, thank you, or apologizing, even though these rules of courtesy are included in its programming.
You can also add a series of notes about the context of who is going to read it or Who is the response addressed to?. You can ask them to write a text for you “as if it were for an 11-year-old child”, “as if they had no idea about the subject” or “as if it were for an expert on the subject”. In this way, it will adjust its level to our interests.
