For decades, we have digitalised education with promises of better teaching, greater access to knowledge, and a generation better prepared for the future. At the same time, criticism is growing over what screens are actually doing to children's concentration, memory, and ability to think independently. Cognitive neuroscientist Jared Cooney Horvath goes so far as to argue that digitalisation has contributed to a historic decline in young people's cognitive abilities. But what happens when we compare the digitalised classroom with children who are educated at home, often in smaller settings, with individual guidance and greater freedom to develop at their own pace? Perhaps the real question is not where children go to school, but why we have spent so long confusing modern technology with better learning.
The Digital Revolution Nobody Really Evaluated
There is something almost absurd about how readily we accepted the digitalisation of education. For generations, we built school systems around teachers, books, writing, conversation, repetition, and sometimes rather demanding intellectual work. Then came computers, tablets, digital learning materials, and permanently connected classrooms. Suddenly, a school that failed to digitalise quickly enough was considered outdated. Computers became symbols of the future, almost regardless of what students actually learned from them. Politicians could have their photographs taken in front of newly purchased screens, school administrators could present ambitious digital strategies, and technology companies could sell their products under the virtually untouchable concept of educational innovation. The question of what happened to children's ability to concentrate, reason, and understand often seemed secondary.
The American-Australian cognitive neuroscientist and former teacher Jared Cooney Horvath challenges precisely this development. In a widely circulated presentation about technology and children's learning, he argues that Generation Z is the first modern generation to demonstrate declining cognitive performance across a broad range of abilities compared with its predecessors. He points to attention, memory, reading comprehension, and mathematics, suggesting that a decisive shift occurred when digital technology began to dominate education. His most provocative argument is that neither the size of the screen nor who paid for it ultimately matters. A computer does not automatically become educationally valuable simply because a school owns it.
Horvath is right to question the almost sacred status of digitalisation. But his conclusion that screens generally damage learning goes further than the available research supports. The OECD's PISA 2022 study found that approximately 30 percent of students across participating OECD countries reported being distracted by digital devices during most or all mathematics lessons. At the same time, students who used digital tools moderately tended to perform better than those who did not use them at all, even after certain statistical adjustments. The problem therefore appears not to be technology itself, but how it is used, how frequently, and what it replaces.
Yet this raises a question that rarely receives much attention in the debate. If the digitalisation of conventional schooling has genuinely created new problems, what might we learn from children who are partly or entirely outside that system?
Homeschooled Children – An Inconvenient Comparison
In the United States, particularly, homeschooling has become an established educational alternative in which parents assume primary responsibility for their children's education. Methods vary enormously. Some families follow structured curricula and administer regular examinations, while others work through projects involving literature, science, practical activities, and discussion. Still others practise forms of so-called unschooling, allowing children's own interests to direct a larger part of their education. Homeschooling is therefore not a single educational model, and that distinction matters. One family may provide exceptionally ambitious teaching, while another may lack the resources or knowledge required to offer an adequate education.
Nevertheless, the research findings are interesting. A systematic research review published in 2026 by researchers associated with the National Home Education Research Institute reported that homeschooled students performed better than comparison groups in 62 percent of the academic-outcome studies included in the review. That sounds impressive, but the organisation is favourable towards homeschooling, and the research field suffers from substantial methodological limitations. Families choosing homeschooling often differ from other families in parental involvement, educational background, financial resources, and priorities. Consequently, we cannot simply conclude that homeschooling itself causes better academic results.
This is an essential qualification, but it does not make the comparison less interesting. Quite the opposite. What if some of the advantages attributed to homeschooling actually reflect something more fundamental: an adult who has sufficient time to pay attention to a child's individual development? In an ordinary classroom, a teacher must simultaneously meet the needs of twenty or thirty students, follow curricula, document results, and manage a social environment in which education constantly competes with other activities. At home, a child may receive more uninterrupted attention, personal feedback, and the opportunity to continue working on a subject until it is genuinely understood.
It reminds us of a simple truth that risks disappearing beneath endless discussions about digital learning platforms: education is fundamentally a human relationship. It is not merely the transfer of information, but the process through which a more experienced person helps a less experienced person develop the ability to understand the world independently.
And this is precisely where the possibilities of homeschooling intersect with Horvath's criticism of screen-based education.
When Learning Is Allowed to Take Time Instead of Filling a Timetable
Perhaps the most interesting aspect of homeschooling is not that it takes place at home, but that it can break away from the industrial organisation of conventional education. The modern school is, in many respects, a product of a society that needed to educate large populations in a relatively standardised manner. Children are divided according to birth year, placed in classrooms, assigned timetables, and expected to reach predetermined objectives within roughly the same period. Historically, this has been an enormous achievement. Universal education has contributed to literacy, social mobility, and democratic participation. But its organisation also involves compromises. A student who needs two weeks to understand a mathematical concept cannot always have that time. Another who understands it after twenty minutes must often wait for everyone else. The system is designed to manage groups, not necessarily to allow every individual to develop at an optimal pace.
Homeschooling, when it works well, can offer something different. An entire morning might be devoted to a historical problem that raises further questions about economics, politics, and technology. A walk through the countryside might become a biology lesson. An unsuccessful experiment can continue the following day instead of ending when the bell rings. Children can discover connections between subjects, something that becomes more difficult when knowledge is divided into separate lessons and rigid academic categories. There is, of course, nothing preventing conventional schools from teaching in similar ways, but their organisational structures often make this more complicated.
At the same time, romanticising homeschooling would be naive. Not every parent possesses the necessary teaching skills, subject knowledge, financial resources, or ability to give up working hours. Some children require specialist support that professionally trained educators are better equipped to provide. Others benefit enormously from a social and intellectual environment where they regularly encounter people with backgrounds and opinions very different from those within their own families. A good school can offer precisely that. Homeschooling can also result in isolation, limited perspectives, and inadequate academic development if the quality of instruction is poor. Researcher Daniel Hamlin observed in a 2024 review that homeschooling research continues to suffer from small samples, selection problems, and a shortage of long-term studies. In other words, we know less than either enthusiastic supporters or determined opponents sometimes suggest.
But that does not change the underlying question. If homeschooled children can achieve strong educational outcomes under certain conditions without the same institutional structure, shouldn't we investigate which elements of their learning environments actually work and attempt to incorporate those lessons into conventional education?
From Independent Thinking to Ready-Made Answers – AI Changes Everything
What makes Horvath's warning particularly interesting is that it arrives just as we face a technological transformation potentially far greater than the introduction of tablets. Artificial intelligence may soon become as ordinary a part of education as the textbook once was. The difference is that AI does not merely present information. It can construct arguments, solve problems, summarise literature, create illustrations, and write entire essays. This threatens to shift the boundary between helping students learn and simply performing their work for them.
Imagine two children investigating why the Roman Empire collapsed. One is asked to research the subject, read different historical explanations, discuss contradictions, and formulate an independent conclusion. The other uses an AI service that produces a beautifully written summary of the main explanations within seconds. Both may submit impressive assignments. But have they undergone the same intellectual development? Hardly, at least not if the second child simply accepts the generated answer. Much of genuine education lies in the encounter with difficulty, in the effort to understand something that is not immediately obvious.
Here, homeschooling may offer an interesting educational experiment. In an environment where learning is built around continuous conversations between children and adults, it becomes harder to conceal a lack of understanding behind a polished digital answer. A parent or tutor following a child's reasoning can ask the next question: Why do you believe that? What evidence contradicts your conclusion? How would you explain this to someone else? Naturally, this is not exclusive to homeschooling. A skilled teacher does exactly the same thing. But individual feedback becomes particularly valuable when machines can produce almost any written assignment imaginable.
The paradox is that AI could simultaneously make qualified, personalised education more widely available. A student without academic support at home could receive explanations, exercises, and guidance at any hour. Technology could strengthen both classroom education and homeschooling. But we must be exceptionally clear about the role we want it to play. Should AI function as a coach developing the student's own abilities, or as an intellectual assistant making the student's effort unnecessary? That distinction may become decisive for the future of education.
Who Actually Owns a Child's Intellectual Development?
There is also a wider social question behind the homeschooling debate, one that reaches far beyond the choice between the kitchen table and the classroom. How much influence should governments, educational institutions, families, and technology companies have over the way children learn to understand the world? In Sweden, opportunities for homeschooling are extremely limited under compulsory education legislation, while other countries have adopted considerably more flexible arrangements. These differences reflect competing ideas about the purpose and responsibility of education. Should schools primarily guarantee every child equal opportunities and exposure to the wider community, or should families be free to choose entirely different routes towards comparable educational outcomes?
There are legitimate arguments on both sides. A society must be able to ensure that children receive basic knowledge, protection, and opportunities to develop independent opinions. But that same society must also be willing to examine its own institutions critically. If schools fail to develop the concentration, reading comprehension, and intellectual independence that education is supposed to provide, the answer cannot always be another digital reform, another learning platform, or another layer of administrative measurement.
UNESCO's 2023 report Technology in Education warned about the lack of robust evidence supporting many of the claimed benefits of educational technologies and emphasised that technology must serve educational needs. This should be an obvious principle, yet it has proved remarkably difficult to maintain when digitalisation has simultaneously become a political objective and a lucrative commercial market.
And this is where I believe Horvath's argument becomes most significant, despite his exaggerations. We should not begin by asking which technologies the next generation needs to use. We should begin by asking which abilities the next generation needs to develop. The ability to read a lengthy text without constant distraction. To understand complicated relationships. To distinguish facts from manipulation. To tolerate uncertainty. To formulate an original thought that does not already exist, neatly packaged, on a screen.
Perhaps the comparison with homeschooled children suggests that the school of the future does not need to be either entirely digital or entirely analogue. Above all, it needs to become more human. Less preoccupied with administering learning and more committed to actually making it happen. More time for conversation, concentration, experimentation, and intellectual curiosity. Technology should be used when it genuinely adds value, not merely because it happens to be available.
It would be a remarkable historical irony if, after investing enormous resources in digitalising education, we discovered that some of the most important educational tools were still a knowledgeable adult, a curious child, a book, and enough time.
And perhaps that is where the most uncomfortable question lies. Not whether homeschooled children are generally more intelligent than others – research has not established that – but whether we have constructed an educational system that sometimes makes it unnecessarily difficult for children to develop the very qualities on which human learning has always depended.
We have spent decades teaching children how to use computers. Perhaps it is time we devoted considerably more effort to ensuring they can also think without them.
By Chris...