Since the end of 2022, the massive influx of generative artificial intelligence tools has profoundly altered daily school life. The launch of ChatGPT by OpenAI marked a rapid and dramatic turning point. In just a few months, this chatbot became second nature for a large proportion of French students, particularly in high school, where its use for homework and presentations has become widespread. Today, AI is no longer a technological curiosity but a learning companion, sometimes discreet, sometimes intrusive. The figures reflect this normalization. A very large majority of tenth-grade students report regularly using ChatGPT for homework. This rapid adoption caught the educational system off guard, even though scientific studies on the real effects of these tools remain limited. Initial international research, however, paints a mixed picture. Some research suggests that AI can facilitate access to complex knowledge, while other research points to a loss of autonomy and a weakening of reasoning when these tools replace personal intellectual effort.
Between increased efficiency and cognitive impoverishment
Recent publications have highlighted a reduction in cognitive load when students use language models for their research, but also a decline in the quality of the reasoning produced. In short, the answer arrives faster, but the reflection process diminishes. For teachers, these effects are no longer merely theoretical. In class, homework assignments sometimes contain glaring inconsistencies, inappropriate vocabulary, or reasoning poorly understood by the students themselves. Faced with this situation, the educational community is trying to adapt. Some teachers observe that AI is being used as an intellectual crutch, with presentations recited without real mastery of the content. Others, however, believe that supervised use can be beneficial, particularly when the tool is introduced after initial independent work, to improve a text or verify an argument. In this case, AI acts more as an enhanced proofreader than as a substitute for learning. The educational system has acknowledged this shift. The French Ministry of Education has published a framework for the use of AI in order to establish clear guidelines. Starting in 2026, secondary school students will follow an awareness-raising program on the PIX platform, designed to develop an understanding of how algorithms work, the ethical issues involved, and the limitations of these technologies.
To train without abandoning critical thinking
From the teachers' perspective, adaptation also involves questioning teaching practices. Abroad, as in France, many teachers have reduced homework and strengthened in-class activities to better assess students' actual reasoning. Data from academic research, particularly in the United States and Northern Europe, confirms this trend toward more oral, reasoned, and less easily automated exercises. In France, a majority of teachers express a need for training in these new tools. A national plan exists, but its implementation is considered uneven. Some teachers rely more on their own technological monitoring to integrate, or conversely, limit the use of AI in their lessons. The main concern remains that of passive dependence, which would weaken students' ability to read long texts, structure their thoughts, and persevere in the face of difficulty. In this context, teacher groups are advocating for a more reasoned use of digital technology, where AI becomes a subject of instruction rather than simply a tool. The goal is not to ban these technologies, but to allow students to do without them when necessary. Some French initiatives are attempting to provide a structured alternative, with educational AI trained on reliable content and designed to guide without doing the work for the student. In schools, artificial intelligence is no longer a hypothetical future scenario. It is becoming a permanent reality, forcing teachers and students to redefine the boundaries between assistance, learning, and intellectual autonomy.
