Carlos Rosique | Alicante, Feb 5 (EFE) that “the way of evaluating and also teaching” is changing in university centers and institutes.
Oliver, scientific director and co-founder of the Ellis Foundation, stressed in an interview with EFE that “the potential of artificial intelligence and specifically of these systems such as ChatGPT must be taken advantage of to help write better and even improve the creative process.”
Potential of ChatGPT and artificial intelligence
Thus, he stresses that the potential and advantages of ChatGPT, for example, are much greater than its drawbacks, and he does not share the decision of the city of New York, which has prohibited its use in public schools, or that of the Institute for Political Studies. from Paris, one of the main training centers for French political elites, which has followed in the footsteps of the American city.
These new programs imply “a deeper transformation” of education, Oliver points out, beyond the prohibition of their use or not, and “it is important to consider what positive elements these great advances bring to ensure that each student performs their potential better, to encourage more writing and to help creativity.
“It’s an extremely powerful tool in the creative context.” In his opinion, when you don’t know how to express something and you type a couple of words into one of these systems, it helps you get out of the block”, Oliver emphasizes. He maintains that teachers should not worry about the appearance of these programs since there are also others that detect whether or not the texts have been written with AI.
They have not been trained to be truthful
Oliver points out that these systems “have not been trained with an explicit veracity criterion” and that, “no matter how well written what they are writing is, it is not necessarily true and it is not necessarily truthful”, since “what the program is to answer correctly what you ask, even if what you ask is not true.
For this reason, he affirms that this “can be a very valuable tool to teach students to question the sources and to understand the difference between what is true and what is not, to learn to contrast the information.”
The woman from Alicante, a tenured academic at the Royal Academy of Engineering of Spain, explains that ChatGPT has 175,000 million parameters that have been trained with billions of texts, most of which are found on the Internet, “and it is difficult, with those volumes of data, someone write down what is true and what is not» to correct it later.
The system “does not understand what it is writing”, since it is a mechanism that “has learned which are the most probable words that should go after others based on that huge corpus, not from the texts with which it has been trained, and That is why it is important not to confuse the one who knows how to write very well with the one who really knows what he is talking about.
Rise of artificial intelligence
The scientist from Alicante also stresses that AI is a science founded already in the 50s, although it seems much closer, and in that time “it has had boom periods that are usually followed by periods of decline, because the expectations that are generated are not comply”.
However, “now there is more talk of a constant spring”, since, beyond ChatGPT, applications such as Krisp have emerged, which eliminates background voices in calls; Beatoven, which creates unique and free music, or other applications that allow you to create completely new images from two or three words.
For this reason, it highlights that this family of “generative models”, that is, these programs that are “an engine of creativity” based on the information that is given to them, “are not only being used in the audiovisual context, but also They do it in chemistry or biology, a sector in which they are helping to synthesize proteins.
part of a revolution
“The concept of using artificial intelligence to generate is forming part of a revolution” that will give “great advances” in the coming years, insists the scientist from Alicante.
“This – the generative model that ChatGPT is included in – is just one type of model,” he explains. «Then we have all the models that recognize speech, that allow translation from one language to another, that allow tumors to be detected in X-rays or that predict what happens in the future in the weather and that help photovoltaics. This is all thanks to AI,” concludes Oliver. EFE