While National Education announces the creation of an Education Anticipation Committee and the upcoming generalization of a course dedicated to artificial intelligence, one observation is clear: AI has already penetrated the daily lives of students. Massively, discreetly, and without waiting for institutional approval.
Generative AI tools are used to reformulate courses, summarize texts, produce homework or structure reflection. This reality, now well documented, is neither anticipation nor foresight: it is imposed on schools and higher education as they are today.
Recent studies converge on one point: students do not primarily use AI to deepen their understanding or improve their skills, but to save time. The tool becomes an accelerator, sometimes at the cost of impoverishing intellectual effort. The risk is not so much that of cheating as that of a progressive shift: confusing production and learning, efficiency and real mastery of knowledge.
Faced with this situation, two reflexes oppose each other.
The first consists of wanting to prohibit, control, sanction.
The second, more demanding, but also more realistic, consists of organizing its use, defining its limits and educational purposes.
The total ban is an illusion. AI is already there, accessible, integrated into practices. Seeking to ban it would amount to moving the problem without resolving it, while widening an additional gap between the real uses of students and the school environment. This approach would also be detrimental in terms of international competition.
Several countries have made the opposite choice. In Estonia, for example, introduction to computer thinking, understanding algorithms and the uses of digital technology is integrated from primary school. More broadly, in many Nordic or Anglo-Saxon educational systems, and even in China, AI is approached as a skill to be mastered, not as a threat to be contained.
Refusing to equip students with these technologies would amount to training a generation out of step with emerging international standards. The question is therefore not whether AI should be used, but how and for what educational purpose.
A profound questioning of evaluation methods
One of the most structuring effects of AI concerns evaluation. Work carried out at home is becoming more and more difficult to assess: not because students commit more fraud, but because part of the production can be assisted, transformed, optimized by automated tools.
This development will probably lead to a strong return to written tests on the table, reasoning in limited time and personal demonstrations. Not out of nostalgia for an old model, but out of educational necessity. When the process is difficult to observe, the result becomes central again.
This move has a consequence that is often underestimated: it restores responsibility to students. Less surveillance, less permanent suspicion, but increased demands on the quality of reasoning and the ability to mobilize knowledge independently.
The decisive role of EdTech
In this context, EdTech occupy a pivotal position. They can either contribute to a workaround logic, produce faster, with less effort, or assume a clear educational responsibility: helping students learn better with AI, without replacing it.
Used intelligently, AI can structure a thought, clarify reasoning, facilitate synthesis, or prepare a demonstration, in short, save time!. Misused, it becomes a crutch that permanently weakens the capacity for analysis and autonomy.
Get out of the false debate
AI acts as a revealer of the fragilities of the educational system: confusion between restitution and understanding, difficulty in evaluating real skills, sometimes excessive emphasis on volume rather than mastery. Wanting to address these issues through a ban would be a simplistic response to a structural problem.
AI will not necessarily impoverish minds. But it will not raise them without a lucid and demanding collective choice either. Recent initiatives are moving in the right direction, provided they are not limited to a symbolic or defensive response.
The real challenge is now educational: training students capable of understanding these tools, of fully mastering their use and of grasping their limits. Not to follow a trend, but to avoid a lasting disconnect between the school, the real practices of students and emerging international standards.
Training students capable of thinking with AI, but also without it, requires a clear framework, accepted uses and renewed demands. It is an educational issue, but also a social and democratic one.
