“A wake-up call” in the education sector: for exam correctors, recognizing a copy written by AI is a challenge

"A wake-up call" in the education sector: for exam correctors, recognizing a copy written by AI is a challenge

A British study claims that proofreaders are often unable to differentiate a “real” copy from one generated by AI. JohnnyGreig/Getty Images

Academia is mixed on the use of ChatGPT, which some believe makes cheating easier. Especially since teachers do not always recognize the productions of this artificial intelligence tool, as shown by a study published in the journal PLOS One.

The authors of this study had proof of this after using ChatGPT to answer around sixty evaluation questions, relating to five modules taught as part of the psychology degree at the University of Reading (UK). -United). The chatbot had to produce different types of responses: short texts and longer ones.

At the same time, undergraduate psychology students had to answer the same questions, but at home. They were allowed to consult their notes and books to help themselves, but they were strictly prohibited from using ChatGPT for this exam.

Exam answers generated by artificial intelligence represented on average 5% of papers, and were submitted to examiners. The latter did not know that, among the copies they had to grade, there were documents written by ChatGPT.

The researchers found that the spellers were, for the most part, unable to differentiate between a "real" copy of another generated by AI. Only 6% of the copies produced by ChatGPT aroused the examiners' suspicions. Some even obtained better grades than those written by undergraduate psychology students.

"A warning shot" for the educational environment

Professor Peter Scarfe hopes the findings of this study, which he co-authored with Kelly Watcham, Alasdair Clarke and Etienne Roesch, will serve as a "wake-up call" for higher education.

"Many universities have moved away from traditional exams in favor of more inclusive assessment methods. Our study shows that understanding how AI will affect the integrity of university exams is of critical international importance. This does not mean that we should return to handwritten exams, but it is necessary for the education sector to evolve to cope with AI" , he said in a statement.

Several academics around the world have taken steps to regulate the use of AI. Some have banned it, while others, such as the Russell Group higher education institutions, view this technology more as an educational tool.

But generally speaking, the world of education has not yet addressed this issue, to the point that UNESCO has published a guide on the ;generative artificial intelligence in education in 2023. Through this document, the United Nations called on governments to "rapidly regulate" the use of AI in schools, so that this technology is "a tremendous opportunity for human development", and not "the source of damage and harm".

Add a Comment

Your email address will not be published. Required fields are marked *

(function(d,s){d.getElementById("licnt2061").src= "https://counter.yadro.ru/hit?t44.6;r"+escape(d.referrer)+ ((typeof(s)=="undefined")?"":";s"+s.width+"*"+s.height+"*"+ (s.colorDepth?s.colorDepth:s.pixelDepth))+";u"+escape(d.URL)+ ";h"+escape(d.title.substring(0,150))+";"+Math.random()}) (document,screen)