The ChatGPT artificial intelligence, capable of writing texts in response to simple questions, has spread massively in the educational world, leading teachers to wonder about the opportunity to ban it or benefit from it.
Since mid-December, just a few weeks after the instrument was supplied by the Californian start-up OpenAI, eight Australian universities announced that they will modify the exams and consider that the use of Artificial Intelligence by students is related to cheating.
In 2023, your tests will now be "guarded" with "an increasing recourse to paper and pen"indicated the leader of the "group of eight" Vicki Thomson, quoted in The Australian newspaper blog.
Recently, after various media outlets referred to the growing use of the tool by students around the world, especially encouraged by TikTok videos, New York public schools restricted access to ChatGPT on their networks and terminals.
The instrument "does not facilitate the development of critical thinking and problem solving skills, which are essential for success in school and throughout life"says Jenna Lyle, spokeswoman for the US city’s education department, in a statement to AFP.
ChatGPT is a conversational robot that was "trained" thanks to phenomenal amounts of data obtained on the net and can "predict" the probable continuation of a text. But although it does not reason, it produces an impressive mixture of correct answers and factual or logical errors, more or less difficult to decipher.
For example, he cites the whale shark (a fish) among marine mammals, he is wrong about the size of the countries of Central America, "forget" some historical events such as the battle of Amiens in 1870 or make up bibliographical references.
In the educational world, some voices speak out against this innovation in teaching methods.
"ChatGPT is an important innovation, but no more than that of calculators or text editors"who eventually got a place at the school, explains to AFP Antonio Casili, a professor at the Polytechnic Institute of Paris and author of "waiting for the robots" (Seuil).
According to him, "ChatGPT can help to make a first draft when you are in front of the empty page, but then you have to rewrite everything and give it a style".
– Ban would encourage use –
The expert also points out that ChatGPT partly shakes up the philosophy of teaching, based on the teacher who asks questions.
This time, the student interrogates the machine and "It is an opportunity for us to see how the students carry out the tasks that we entrust to them, to make them work on the fact-checking, and to verify if the bibliographical references generated are correct."Casili analyzes.
For Olivier Ertzscheid, a researcher at the University of Nantes (western France) in information sciences, the ban on the instrument is in any case "counterproductive" as it reinforces the desire of students to use it.
As after the advent of Wikipedia or search engines, the option for teachers is, according to him, "experience the limits" of those instruments.
Finally, responses are being generated to detect the texts generated by Artificial Intelligence.
The online service GPTZero prepares, for example, an offer dedicated to education professionals and OpenAI works on a "statistical watermark" placed when the text is generated. In other words, cheaters are already warned.