AI Can't Fix the Student-Motivation Problem

In 2023, Sal Khan — the founder of Khan Academy — promised that artificial intelligence would bring about the greatest positive revolution in the history of education. A personal tutor for every student on the planet, 24 hours a day, in any subject. But just a year later, Khan himself admitted that the launch of Khanmigo, his company's educational chatbot, was "a non-event" for most kids. Although access has grown from 40,000 students to almost a million in a few months, the truth is that almost no one really uses it. And here lies the twist: AI can offer tailored exercises, personalized explanations, brilliant animations — but it can't create the desire to learn. No algorithm, no matter how sophisticated, can push a student to put in the effort if they don't see the point. Kristen DiCerbo, head of learning at Khan Academy, puts it bluntly: “Learning is mental effort, and it's often not fun. How do we make kids want to tackle it?” Artificial intelligence cannot fill this motivational void. The problem is not access, it is desire. A key fact: only one-third of American students are truly engaged in school, and this share has remained stable for years — despite all the new digital tools. And even among these few motivated students, only 5% really use platforms like Khanmigo enough to see the benefits. It's the "five percent rule," as Laurence Holt calls it: most use little or nothing, and those who were already ahead go even faster. The risk? That technology widens the gap instead of closing it. Sal Khan himself, after years of enthusiasm for AI, now admits that the real leverage is investing in people. Because education, after all, remains a social phenomenon. Ron Ferguson of Harvard has studied thousands of classes: the teachers who really motivate are those who know how to create collaboration and discussion among students. Kids learn more when they discuss, challenge each other, and see that their classmates are also struggling. The same logic applies to gyms: those who hate exercise work out more in a group than alone. But there is a paradox: while the best teachers know how to use group dynamics to motivate, they are few and difficult to train. And public schools often choose teachers based solely on their degrees and theoretical knowledge, not on how well they can really spark curiosity or hold the class's attention. Ferguson found that teachers who push students to work hard and aim high have two characteristics: they are engaging — they manage to bring even the most difficult subject to life — and they make every student feel that their success really matters. These are not innate skills: they are learned through experience, feedback, and by observing the best colleagues. Roland Fryer of Harvard confirms this: coaching and peer observations help teachers grow more than any theoretical course. And Mary Burns, a former teacher and now a researcher, recalls that during the pandemic, we saw what happens when the value of the group is underestimated: the loss of learning was also a psychological wound. Many kids, simply because they were isolated, stopped trying. The aspect that remains outside the tech debate is that kids, after all, don't learn for AI, but for people: to impress a teacher, to collaborate or compete with friends. Justin Reich of MIT puts it clearly: “They care about people, not the machine.” And it's not enough to make bots smarter or more scalable, as long as no one has found a way to make them a source of motivation. If you think an algorithm will be enough to change schools, you're on the wrong path. AI cannot solve the problem of motivation, because motivation comes from human relationships. If this story has made you see school in a different light, you can report it on Lara Notes with I'm In — it lets you say that this idea now really concerns you. And if you happen to discuss it with someone, you can tag them with Shared Offline: it's the way to save a conversation that has struck you, so it doesn't get lost. This story comes from The Atlantic and saves you 1 minute compared to the original article.