Researchers from the Federal Polytechnic School of Lausanne (EPFL), one of the most prestigious universities in Europe, subjected the ChatGPT artificial intelligence (AI) systems to tests from various studies of the institution, observing that they correctly answer most of the questions .
The ChatGPT 4 and 3.5 language models answered around 65% of the questions correctly using immediate strategies and basic data, but with increased knowledge they can achieve up to 85% correct answers, a surprising result according to scientist Anna Sotnikova, from the EPFL Language Processing Laboratory, who led the study.
These AI language models were subjected to 50 EPFL courses at various levels, from degrees to masters or online programs, around disciplines ranging from computer science to mathematics, biology, chemistry or physics, among others.
We recommend: AI has the greatest impact on the benefit of urban regions: OECD
AI could weaken college students’ learning
“If the study were repeated today the results would be even better,” the researchers warned, taking into account the continuous advances of this technology.
The study sought above all to examine the effects that the use of AI can have on higher education students, and in that sense it did not hide the risk that the use of these language models affects the educational process of university students necessary to increase their academic abilities.
“There is a fear that if the models are as capable as they seem, students will use them to take ‘shortcuts’ in the process necessary to learn new concepts, which could lead to weaker knowledge bases that will make it difficult to future absorb more knowledge.” complex,” warned those responsible for the study.
With information from EFE
Do you like photos and news? Follow us on our Instagram
