ChatGPT: “Will artificial intelligence replace teachers?”

Based on its program and on millions of pages of textual data in different languages ​​(newspaper articles, but also all of the Latin works of Saint Augustine, or even works on the history of Japan and cookery manuals, which I was able to ascertain by asking questions), Chat GPT produces impressive results, which must have been tested to measure their novelty. The public availability of this tool (for now free) will disrupt our habits, not in the months to come, but in the days that follow, when the media tidal wave, formed today on specialized sites, will have swept over the shores of the general public. The consequences must be measured on a case-by-case basis, profession by profession, as concretely as possible, and first of all in education. Because our job is to cultivate natural intelligences, at a time when they enter into competition or synergy with machines.

Since ChatGPT can not only answer questions, but also answer them “in a certain way”, the smart student can already ask ChatGPT to produce a text for him.

Julien Auriach

Finish the homework! The immediate consequence of the arrival of ChatGPT is the end of graded homework, what we call in our jargon: DM. The same goes for presentations. It had already become difficult for five or ten years, to give these exercises. In question ? The reflex of some students to go to Google for the answer to the subject of an essay. Obviously, rewriting a text to the letter was easily spotted, which is why the students, naturally gifted in paraphrasing, spent hours disguising their borrowing. The whole paradox was then that plagiarism was still a work of personal synthesis, sometimes longer than that which would have consisted in not cheating. As for the less gifted, tools such as www.paraphrase.io have arisen in recent years. They allow, via an algorithm, to reformulate a text automatically to deceive the vigilance of a bad-tempered teacher, of the kind who types the work of a student on Google, to check its originality. Plagiarism, we read on Wikipedia, “consists in copying an author, […] without quoting or saying so […]. It is often equated with immaterial theft.” Reformulated with the software, this gives: “Plagiarism is copying an author without citation or acknowledgment. It is often likened to petty theft.” The arrival of ChatGPT, by mixing google and paraphrase.io, and pushing them further, definitely ruins the very principle of DM.

Since ChatGPT can not only answer questions, but also answer them "in a certain way", the smart student can already ask ChatGPT to produce a text of three sentences by adding a detail of his course (to make it true) and do it "in the style of a ninth grader", for example, about the plagiarism. This is undetectable and in the context of our example it gives the following: "Plagiarism is when you copy someone else's work without giving them credit. If you do that, you risk being punished by your teacher or even by your school. Last year, I heard of a student who was accused of plagiarism because he copied another student's math assignment without telling him.»

Dear teacher friends, stop the DMs. It's no use anymore. This does not mean that we should stop working at home, far from it. Personal learning and confrontation with texts and information to be learned in view of a homework assignment is important. But the note must be developed via a supervised assignment, during which you can ensure the absence of artificial intelligences, and therefore the sole presence of natural intelligences. If the former can be trained, only the latter can be cultivated, and that is our job.

For ChatGPT's answers to be relevant, the questions must be relevant, and how would they be, without a fertile ground for them to arise?

Julien Auriach

Technological progress must lead to pedagogical progress. This progress pushes us to a pedagogical step forward, which will perhaps appear to more than one dreamer as a step backwards, but it is not so. It is urgent to drive the screens out of the school. We could once delude ourselves, believing that digital at school was necessary. And until recently: the department of Seine-Maritime has just made availablees iPad for college students sixth grade, ignoring the discoveries of science, synthesized among others by Michel Desmurget in his work, The Digital Cretin Factory (2019). This is culpable negligence, which should one day be able to be dealt with in court, if there were something like a charge of “negligently endangering the [intellectual] life of another”.

The arrival of ChatGPT, the first reliable device for developing text according to instructions given in natural language, is an opportunity for National Education to revise its relationship to digital technology. Mastering AI is a challenge that every student will have to achieve in their field during their post-baccalaureate studies. But middle and high school students must first develop their natural intelligence, through exercise and study, organic and without inputs. Hear without a screen. This, in addition to being the condition of their thinking and of their freedom as women and men to come, is also the condition of their effective use of the AIs of tomorrow. Because for ChatGPT's answers to be relevant, the questions must be relevant, and how would they be, without a fertile ground for them to arise? The rise of artificial intelligences puts natural intelligences before a choice: either bondage from an early age, or mastery at a more advanced age, permitted only by growth independent of the one and the other. Consider, then, this sentence from Cicero: "As a fertile and well-cultivated field yields good and abundant fruit, so the human mind cultivated and nourished with knowledge produces beautiful and useful thoughts." (Cicero, Tusculans, I, XV, translation by ChatGPT).