A Tutor for Every Student Is Finally Plausible. The Teacher Still Matters More.

One of the oldest findings in education is that individual tutoring works remarkably well — a learner with a patient tutor who adapts to their pace, notices exactly where they are stuck and explains a concept until it clicks, tends to outperform the same learner in a class of thirty. The obstacle has always been that human tutors do not scale; there are not enough of them and not enough hours. AI tutoring is compelling because it offers a version of that individual attention at a scale and cost that human tutoring never could — a system that will explain a topic a dozen different ways without impatience, meet a student at their actual level, and be available at eleven at night when the homework is due. The genuine advance is in personalisation. A classroom teacher, however skilled, must pitch a lesson to the middle of the room, leaving the students ahead of it bored and the students behind it lost. An adaptive tutor can, in principle, move at each learner's pace: slowing down and finding a new explanation when someone is struggling, moving briskly and offering harder material when someone has grasped it, and doing so without the student feeling singled out or embarrassed to ask the same question a fourth time. For a learner who has fallen behind and needs patient repetition, or one racing ahead who needs to be stretched, the difference between a lesson aimed at the average and one aimed at them can be the difference between disengaging and thriving. The important caution is the same one that applies to AI everywhere, and it matters more with children: these systems can be confidently wrong, and a tutor that teaches something incorrect with total authority is worse than no tutor for a learner who cannot yet tell the difference. There are also things a piece of software does not do — notice that a child is struggling for reasons that have nothing to do with the material, provide the encouragement and relationship that motivate a discouraged student, teach the judgement and social learning that happen between people, and model what it is to think through a hard problem rather than just receive an answer. A tutor that hands over answers too readily can even undermine the productive struggle that learning requires. The realistic and genuinely hopeful picture is AI tutoring as a powerful tool in a teacher's hands rather than a substitute for the teacher. Used well, it takes on the part it is good at — endless patient practice, instant personalised explanation, immediate feedback — and frees the human educator for the parts only a human does: motivating, mentoring, spotting the struggle behind the wrong answer, and teaching a child how to think rather than merely what to know. The schools that get the most from this will not be the ones that replace teachers with software to cut costs, but the ones that give teachers a tool that finally makes one-to-one attention scalable, while keeping the human relationship that makes education work at the centre.