Walk into certain well-funded classrooms today and you will see something that looks, at first glance, like the future: every child on their own device, working through their own adaptive pathway, at their own pace, on material calibrated precisely to their level. No child bored, no child left behind. It is a compelling picture, and it has a compelling name. Who, after all, could be against learning that is personalised?
The evidence for it is real but thinner than the marketing suggests. The most cited evaluation is the RAND Corporation's study of schools in the Gates Foundation's personalised learning initiative, which found that students in these schools made gains in maths and reading significantly greater than comparison students over two years, moving from below to above the national average (Pane et al., 2015). Promising — though the sample was mostly charter schools that had volunteered to pioneer the approach, which rather flatters the results. RAND's larger follow-up a few years later was notably more muted, finding no statistically significant effects on average (Pane et al., 2017). This is a familiar pattern in education: a real but modest benefit, inflated by enthusiasm.
But my concern is not really about test scores. It is about what school is for.
The atomisation problem
Strip the idea back to first principles. A school does two jobs at once: it transmits knowledge, and it socialises children into a society. Personalised learning optimises the first job by quietly dismantling the second. If every child follows their own pathway, at their own pace, on their own screen, the shared experience of the classroom — the argument, the negotiation, the group task that goes wrong and has to be rescued — evaporates. What remains is twenty-five individuals with headphones on, each in a private educational bubble.
This matters because the skills that make adults functional are largely learned in the friction of other people. Graceful competition — wanting to win, losing, and behaving decently about it. Teamwork with people you did not choose and would not have chosen. Compromise, patience, the ability to hold a shared standard. These are not extras that schools provide around the edges of the curriculum; for many children, they are the main event. A system that atomises learning risks churning out hyper-individualised, academically serviced adults who are socially dysfunctional — each one perfectly catered for, and none of them practised at living with others.
The research on cooperative learning underlines the point. David and Roger Johnson's decades of work, synthesised in meta-analyses spanning hundreds of studies, found that cooperative learning consistently outperforms competitive and individualistic structures — not only on achievement, but on the quality of relationships between students (Johnson et al., 1981; Johnson & Johnson, 2009). Children working together towards a shared goal learn more and learn to like each other more. The social and the academic are not in tension. Done well, they are the same mechanism.
A fair hearing for the other side
None of this is an argument against personalisation as such. Every good teacher differentiates; the alternative — marching thirty children through identical material at identical speed regardless of need — is its own kind of neglect. Adaptive software can be genuinely useful for drilling foundational skills, for catching the child who has silently fallen three years behind in maths, for freeing the teacher to work with the humans in front of them. The RAND findings, modest as they are, point in a real direction. And the children who benefit most from tailored pacing are often the ones the one-size-fits-all classroom serves worst.
The question is not whether to personalise, but what to personalise and what to keep shared.
My tentative rule, after a few years of doing this: personalise the inputs, socialise the experience. Let a child practise their times tables on an adaptive app at their own level — that is exactly where the technology shines. But the essay, the project, the experiment, the debate — the things that make school a rehearsal for life — should be done together, in the open, with all the productive awkwardness that involves. School is one of the last places where people of different backgrounds and abilities are obliged to cooperate with strangers. A society that loses that will find cohesion much harder to rebuild than a few points of reading growth are to gain.
Personalised learning is a good servant and a poor master. Used as a tool inside a communal classroom, it earns its place. Used as the organising principle of the whole school, it solves the wrong problem — and creates a lonelier one in its place.