Instant answers kill the desire to learn

Between helping with learning and handing over schoolwork to technology, education is looking for the right relationship with AI

6317 views 0 comment(s)
Photo: Reuters
Photo: Reuters
Disclaimer: The translations are mostly done through AI translator and might not be 100% accurate.

You would have to be either very pretentious or a time traveler to know what impact artificial intelligence (AI) will have in a few decades on the way we learn, remember information and think. However, it is worth thinking about it now that AI has become a part of the everyday lives of students. It is important to prepare for the risks that accompany this technological and cognitive upheaval, but also to alleviate the fears that dominate the public debate three years after the emergence of ChatGPT.

Experts in the education system and universities agree on one obvious fact: entire groups of students are now avoiding learning by letting AI do the work for them. Behind this practice lies the human tendency, well-known in educational sciences, to so-called cognitive offloading, or shifting mental effort to external tools. “Every working person is characterized by a certain tendency towards laziness, and this technology further encourages it because it frees us from a whole series of mental operations,” explained Philippe Mérier, emeritus professor of educational sciences at the University of Lyon and lecturer on artificial intelligence.

Even when a student decides to use AI out of curiosity and a desire to develop ideas, the academics Monde spoke to express serious concerns. “The research is quite clear on this point: thinking develops over time, including through the experience of frustration,” stressed Mérier. “What concerns me about AI is speed, because there is a risk that it traps students in a kind of present-moment focus that prevents deeper development. Moreover, an immediate response can satisfy the desire to know something, but at the same time kills the desire to learn.”

Fears are also growing because the consequences of using AI are difficult to discern. Most existing research has involved small groups of students, often placed in simplified or artificial situations. Such research is quickly becoming outdated as the capabilities of generative artificial intelligence (AI) rapidly advance. One question that is causing concern is, for example, how AI affects our ability to remember information. According to the theoretical premises of educational science, the less actively a student is involved in completing a task, the weaker their immediate memory is, and the more difficult it is for them to later recall that information in a different context. Several studies have already confirmed this phenomenon, indicating weaker memory in students who use AI.

Learning partner

But new research paints a somewhat more complex picture. In a study published in 2026 among 14- and 15-year-old students in the United Kingdom, researchers at the University of Cambridge compared how much information students remembered three days after reading a text. One group used traditional note-taking, another was allowed to take notes and use AI, and a third used AI alone, without handwritten notes. Memory was good in both groups that were allowed to take notes, confirming that note-taking is still a great way to remember. However, students who had access to both tools found that AI helped them understand the text more quickly and became more interested in it. “The study demonstrates the benefits of combining traditional learning activities with AI, as opposed to using AI alone,” the researchers concluded.

AI
photo: REUTERS

When it comes to these risks, not all students are in the same position. The way AI could deepen inequalities has begun to worry the education community, especially in France, one of the Organization for Economic Cooperation and Development (OECD) countries where academic performance is among the most strongly linked to parents’ social status. Cédric Nodeau, a researcher in educational sciences, studied, for example, how a small group of 26 high school students from the Paris region used AI, after nearly 500 students completed a quantitative questionnaire.

Three profiles emerged. The first are “engaged and thoughtful” students, with significant academic capital, but also less privileged students who understand digital technologies very well and for whom “AI is used as a tool to structure thinking”. There are also “occasional legalists”, who sometimes turn to AI for research purposes, but their approach is accompanied by “fear of cheating”. Finally, there are “opportunistic students”, who use the tool as a way to “completely leave schoolwork” to AI. “This kind of direct attitude towards the answer, without any critical distance, is often associated with greater school vulnerability”, the researcher concluded.

"I'm concerned about speed with AI, because it risks trapping students in a kind of present-moment focus that prevents deeper development. In addition, an immediate response can satisfy the desire to know something, but at the same time kills the desire to learn," said Merier.

Yet, in a sign that this debate in educational science is only just beginning to heat up, experts are increasingly considering the possibility of AI becoming a true learning partner for any curious and motivated young person, regardless of their social background. Despite his concerns, Merier said AI “opens up extraordinary possibilities.”

“Individualized instruction for each student is a scenario that researchers generally consider highly likely,” he explained, noting that some AI systems are designed to support students in the learning process. “BloomStudy, for example, doesn’t provide ready-made answers; instead, it asks questions to guide learning.”

Tendency to flattery

High school students haven’t waited for someone to show them how AI can be used constructively, as literature professor Jean-Michel Le Beau wrote about in his book “School Before Artificial Intelligence: The Choice of Pedagogical Intelligence.” Some students are turning lessons into guessing games, suspenseful stories, or investigations. They’re asking AI to predict questions that might come up in an oral French exam. On the website Chatbac.fr, they chat with virtual characters inspired by Rimbaud, Colette, or Zola.

At university, AI can help to deepen understanding of a concept by exploring its history, epistemology and critique. If generative AI is asked to review a written assignment, students can receive feedback that is “immediate and constructive, two important prerequisites for effectiveness,” said Christophe Genès, head of the department of educational sciences at the University of Nanterre. “Provided you know how to protect yourself from its tendency to flatter you and confirm pre-formed beliefs,” he warned.

Among optimists, AI is fueling another big hope: that it will save time by taking over simpler intellectual tasks. “This is consistent with what happened during the Renaissance,” said Andre Tricot, a professor of cognitive psychology. “With the advent of the printed book, memorization suddenly became a less important skill, while understanding became more important than memorization,” he explained. In the 20th century, the advent of Internet browsers made access to knowledge even easier, encouraging us to “relocate” some of our memory outside of ourselves.

“AI only answers the questions we ask it, but we ask those questions because we already know something,” warned Merier, emphasizing the importance of not leaving our most precious knowledge to external tools. Tricot also expressed concern: “Nothing fosters belief formation like ignorance,” he warned, “and nothing fosters doubt and questioning like knowledge.”

Attitudes towards AI have yet to mature

All technological revolutions have shaped human intellect. Artificial intelligence could bring about its own changes. As linguist Walter J. Ong pointed out in his 2014 book “Orality and Literacy: The Technology of Words,” before the advent of writing, intelligence was measured by the number of sayings a person knew—truths were expressed concisely and transmitted orally. Writing, as “a deeply internalized technology, built into the very mechanisms of thought,” enabled the expansion of human intelligence and the achievement of levels of abstraction that were previously impossible.

If artificial intelligence is causing such concern, it is also because, by promising to spare students the tedious and time-consuming task of reading, it could further deepen the shift away from books - a trend already fueled by the use of screens - even though it is well documented that the neurological stimulation that reading provides has numerous benefits. Teachers are also most concerned when it comes to literature and humanities subjects, because it is in these areas that students use AI the most. Among the more than 500 high school students who responded to Node's questionnaire, the three subjects in which they used AI the most were French, in first place, followed by economics, and economics and history.

In ChatGPT’s third year, as in the early years of previous technological upheavals, both dangers and opportunities seem to be multiplying. Can the worst and best outcomes happen at the same time? “To a certain extent, technology is neutral,” said Merier. “But the slope is often steeper in one direction than the other. We need to remain vigilant, although it is still too early to know which way will prevail.” Tricot added that it is difficult to see clearly at the moment, partly because the way these tools are used is still far from mature: “We will have to build education around artificial intelligence and see how our relationship with it develops. We do not have enough time distance to assess that yet.”

The psychologist, who has studied technological revolutions, remains optimistic: “The first years of photography or television were a time of great naivety, followed by a phase of cultural maturation,” he said. However, during this period of adaptation, schools will have to develop “counterbalances” to the dangers of AI, warned Mérier: “Homework will have to replace activities such as classroom debates, and students will need to be taught the principle of reversibility: I only use AI for tasks that I already know how to do independently.” It remains to be seen whether national education systems, known for their slow adaptation, will be able to meet this challenge.

Translation: NB

See more: