Vai al contenuto
vati.it

← All reflections

AI doesn’t serve schools — schools serve AI

Valentino Pietrobon · 11/11/2026

AI, a tool for teachers.

The motto is a deliberate reversal. We often hear that artificial intelligence "serves schools", as if bringing it into the classroom were enough to improve learning. I think the opposite is true: it is schools that serve AI, because schools are where people learn to use it well. And AI, before it helps students, can be of enormous help to teachers.

Where AI is not needed: cognitive processes

School develops cognitive processes: remembering and understanding, connecting ideas, analysing a problem and breaking it down, abstracting, arguing a case, evaluating a solution, sticking with a line of reasoning that doesn't work out the first time.

None of this needs AI. It needs students who put their own minds to work. An exercise done by an artificial assistant is an exercise the student has not done: the result is there, the learning is not. Handing over your thinking at the very moment you are building it means giving up on building it.

Where AI matters: competences

But school does not only develop cognitive processes: it also develops competences. The Council of the European Union's Recommendation of 22 May 2018 on key competences for lifelong learning defines them as a combination of knowledge, skills and attitudes: knowledge is made up of facts, concepts, ideas and theories; skills are the ability to carry out processes and use what one knows to achieve results; attitudes are the disposition to act or react to ideas, people and situations.

A competence, then, is not something you possess in the abstract: it shows itself in action, within a context. And in the real contexts of study, work and life, information technologies — and now AI — can play a decisive role in personal success. It is no coincidence that digital competence is one of the eight key competences in the Recommendation, and that the European DigComp 2.2 framework (2022) explicitly includes interacting with artificial intelligence systems. Those who use these tools knowingly have a real advantage; those who use them without understanding them become dependent on them.

That is why the use of AI, like that of ICT, has to be taught:

  • knowing when it makes sense to use it and when it does not;
  • writing precise requests and checking the answers;
  • recognising mistakes, made-up content and bias;
  • respecting your own data and other people's;
  • always remaining responsible for the result.

It is not about banning it, nor about letting anything go, but about teaching a use that is informed and purposeful.

For teachers: the keystone

This is where AI can truly change the teaching profession. For years, educational guidelines have asked us for teaching that is thought through, designed and assessed: competence-based planning, authentic tasks, assessment rubrics, personalised pathways, consistency between objectives, activities and tests. These guidelines are widely shared, but in everyday practice they are hard to apply: their complexity demands time that is often not there.

AI can take on part of this work: drafting a plan to be reviewed, linking competences, skills and knowledge, generating variants of a test, building a rubric from the objectives. The teacher remains the one who decides, chooses and assesses; but can finally spend their time on the choices that matter, rather than on the repetitive side of the job.

This is what I experienced this summer, building tools that start from my own course notes and help me plan lessons, tests and competence-based assessment. They did not teach in my place: they let me teach the way I had always wanted to.

AI does not serve schools as a shortcut for students. It serves teachers, to do better the kind of school we already know we should be doing.

References: Council Recommendation of 22 May 2018 on key competences for lifelong learning (2018/C 189/01), Official Journal of the European Union C 189, 4 June 2018; DigComp 2.2 – The Digital Competence Framework for Citizens, European Commission, 2022.

More reflections

12/11/2026

Zero righe di codice

Tutti i miei programmi sono nati senza che scrivessi una riga di codice. Non perché la programmazione non serva più: proprio il contrario

Leggi →