Reinventing school in the era of generative AI

The rapid integration of generative artificial intelligence (AGI) into educational settings is generating both enthusiasm and concern. While some see it as a pedagogical revolution capable of unleashing cognitive potential, others warn of the risks of excessively outsourcing our intellectual functions. In this complex situation, it is essential to move beyond simplistic contrasts to address the real effects of these tools on learning, attention, motivation, and metacognition.

Generative artificial intelligence is profoundly changing our relationship with motivation. By simplifying content production, it appears to offer a solution to the frequent demotivation linked to creative blocks. Indeed, for some students, the ability to quickly overcome difficult initial phases thanks to AI can be a powerful motivational lever, reducing the fear of failure and thus stimulating the desire to learn. However, this immediate comfort could also foster a reliance on extrinsic motivation: learning not out of intellectual curiosity but simply out of a quest for efficiency and immediate results. This phenomenon raises crucial questions: should schools prioritize short-term cognitive efficiency or value sustained, autonomous effort as a driver of personal development?

Attention is another major focus in analyzing the effects of AGIs. In the digital age, maintaining learners' attention is a constant challenge. Unlike traditional teaching tools, AGIs can personalize information and precisely structure interactions, thus facilitating better attention management among learners with difficulties. However, this automatic structuring also risks preventing students from fully developing their own attentional strategies. The intensive use of AGIs, particularly in the writing field, could thus hinder the construction of an essential learning skill: that of organizing and maintaining attention autonomously and sustainably. The real question, therefore, is how to use these tools in a way that strengthens rather than weakens learners' intrinsic attentional faculties.

Metacognition, the ability to reflect on one's own cognitive processes, is also profoundly impacted by the use of AGIs. On the one hand, these technologies can act as powerful metacognitive facilitators by providing immediate feedback and allowing students to analyze their errors quickly. This potentially encourages a reflective process beneficial to learning. On the other hand, the ease with which AGIs generate responses can significantly reduce the individual metacognitive effort required for analysis and self-assessment. This is referred to as "metacognitive laziness," a phenomenon whereby students excessively delegate their thinking and evaluation skills to machines. The major pedagogical challenge then lies in adequately supporting the use of AGIs, in order to actively develop learners' critical thinking and cognitive autonomy.

Finally, the overall cognitive effects of using AGIs in education are ambivalent. While these tools allow for the effective outsourcing of laborious tasks such as document research or initial writing, thus freeing up valuable cognitive resources for more complex or creative activities. But this outsourcing is not without consequences: by reducing the need for deep cognitive efforts and active appropriation of knowledge, it could weaken the development of fundamental skills such as memorization and critical reasoning. This risk is all the more worrying given that younger generations, accustomed to these tools, could gradually lose the ability to fully engage their cognitive functions without external assistance.

Thus, given the uncertainty inherent in the contrasting effects of generative artificial intelligence, it becomes essential to consider educational governance that is open to the plurality of knowledge and aware of the underlying value issues. Educational policies cannot be satisfied with decisions based solely on empirical studies, however rigorous they may be. On the contrary, they will have to articulate scientific results with a broader democratic reflection, capable of clearly explaining the normative presuppositions on which our educational choices are based.

This democratic approach requires, in particular, integrating sociological and ethnographic perspectives to better understand how students actually use these tools, according to their varied social and institutional contexts. Rather than succumbing to a simplistic view of risks and benefits, the major challenge is to collectively decide what type of education we wish to promote through the integration of AGIs.

In short, the debate surrounding generative artificial intelligence in education must rise above the simple opposition between technological optimism and radical skepticism. It requires a transparent dialogue on the educational values and learning models we wish to defend. More than a simple technological or cognitive issue, it is the very definition of the human and its educational future that is now at stake.

To learn more: read our analysis of French education in the age of AI and discover When AI shakes up the classroom.