Artificial intelligence and school: how the way of studying, testing skills and teaching changes in the ChatGPT era
School is the place where the effects of artificial intelligence become immediately visible in homework, tests, essays, in the way of studying and returning what is requested. Students live in a society that celebrates speed, simplification and automation, and they see adults using AI to write emails, summarize documents, prepare presentations, program and look for ideas. The educational challenge cannot therefore be delegated to the school, because it concerns everyone. Now, what to do? In the short term, it is necessary to contain clumsy, cunning and purely substitutive use of AI, while at the same time training in conscious, explicit and mature use. There is a crucial difference between using a tool to learn better and using it to hurry. If a student asks the machine to write an assigned text, he can obtain a decent result without having done the mental work for which that exercise existed. Writing helps to organize thinking, summarizing forces us to distinguish the essential from the accessory, solving a problem means building a strategy: well, delegating these operations too early means obtaining the product without developing any skills. Furthermore, the joke is sensational: the student in question does not practice – the machine, paradoxically, does (and wants nothing else). The question, however, is not to start a crusade against AI at the end of this article, but rather to understand how to make sure that it is not usable – working on the instructions and rhythms in the classroom – and also, in addition, to investigate how it was used, for which step, with what degree of autonomy and control. A student can submit his own text asking her to identify the weak points of the argument, he can be accompanied in the solution of a problem through subsequent questions, he can compare his own summary with the one proposed by the machine to understand what he has overlooked. He can also use it as an interlocutor, asking for objections, simulating a question, asking for different explanations of the same concept. Let’s be clear, he can very well do without it, without problems and indeed with great advantage for the acquisition of a series of basic skills: of course, this exclusion will not be permanent and sooner or later someone will have to educate him, help him, so as not to leave him to digital self-learning, which could put the critical and conscious use of the tool at risk. Consciously using AI means learning not to trust the first answer, verifying data and sources, recognizing omissions, asking for alternatives, realizing that a very convincing formulation may contain errors. It is a form of literacy that concerns judgment much more than the so-called prompting and which cannot be delegated entirely to the school. Families, universities, businesses, institutions and platforms participate in the construction of the rules with which we will learn to decide what to entrust to the machine, what to declare, what to control and what to continue to do personally. This is the immediate challenge. The deeper one concerns the very way we learn. Artificial intelligence does not just make information available, it also produces texts, interpretations, solutions and summaries. Many tasks that we have used for decades to verify a skill can be performed in a few seconds by a machine. It then becomes necessary to ask ourselves what we really want to measure, without transforming school into a continuous investigative activity in search of what is human and what has been produced by AI. The process will probably count more, therefore knowing how to explain a choice, defend a reasoning, correct a plausible or wrong answer, compare solutions, apply what you know to a new situation. Learning times will also change, because an artificial system can adapt explanations and exercises to the individual student, leaving the teacher with a role that is even more focused on understanding the person, on motivation, on relationships and on the choice of what is worth learning. The relationship with notions will also change. The immediate availability of information does not make knowing useless, because without knowledge it becomes very difficult to judge the quality of what a machine produces. We will have to choose better what needs to be known, what is understood in depth, what is important to know how to do without tools and what we can entrust to them. A first practical response already exists and involves work in the classroom. We will need to get used to having people write more in class, to construct exercises that allow us to see the path, to make school more laboratory-based – a great supporter of frontal lessons writes it here, which retains enormous value when it serves to organize knowledge, provide structures, open perspectives. Alongside that lesson, the time in which students do, test, discuss, correct and show how they arrive at a result will have to increase. Above all, orality will have to become central again. AI could give it back a weight that the school had progressively reduced. In the questioning, in the discussion, in the presentation of a work we see what a person knows and how much he knows how to use what he knows. You see if he understands a question, if he knows how to connect information, if he can handle an objection, if he finds the words to explain a concept, if he knows how to adapt it to a different situation. Orality brings together knowledge, understanding, language, readiness and ability to relate, and makes it much more difficult to confuse a performance produced by the machine with a truly acquired skill. In a world in which producing an acceptable answer becomes increasingly easier, the skills that come before and after the answer will gain weight, from understanding the problem to formulating the question, from judging the result to the ability to choose between alternatives, taking responsibility for the decision. The educational challenge of AI thus ends up being much bigger than AI itself and forces us to return to the fundamental question of what learning really means and what part of cognitive fatigue, just as it becomes possible to avoid it, continues to be indispensable to a person’s education.
