Norway has announced one of the world’s most significant education policy shifts regarding artificial intelligence.
From late August, children aged 6–13 will largely be prohibited from using AI tools in elementary schools. Students aged 14–16 will only be allowed to use AI under direct teacher supervision, while older students aged 17–19 will be taught how to use AI responsibly as preparation for higher education and the workplace. Prime Minister Jonas Gahr Støre has made the reasoning clear: children must first learn the fundamentals of reading, writing, and mathematics before relying on AI assistance.
At first glance, this may seem like a surprisingly cautious move in an era where governments and businesses are racing to adopt AI. But Norway’s decision is part of a much broader trend emerging across Northern Europe.
The Great Educational Technology Reassessment
For more than a decade, schools across the developed world embraced digital transformation. Tablets replaced textbooks. Screens replaced paper. Digital learning platforms became the norm. More recently, AI has been presented as the next educational revolution.
Yet some of the countries that adopted educational technology most enthusiastically are now beginning to reverse course.
Finland is re-evaluating its long-standing commitment to digital-first classrooms. Schools are increasingly reducing screen time and reintroducing physical textbooks and handwritten work amid growing concerns from teachers, parents, and researchers about concentration, distraction, and the cognitive effects of constant digital engagement.
Sweden has also moved away from the “tablet-first” model, investing heavily in new textbooks and reducing the use of digital devices among younger children. Policymakers have expressed concerns that excessive screen exposure may be undermining literacy and foundational learning skills.
What is striking is that these countries are not anti-technology. They remain among the most technologically advanced societies in the world. Their concern is not whether children should use technology, but when and how.
What Developmental Science Tells Us
One of the strongest arguments supporting Norway’s approach comes from cognitive and developmental research.
Children do not learn in the same way as adults.
Before higher-order thinking can develop, children need to build what psychologists call “foundational knowledge structures” — the basic reading, writing, language, and mathematical skills that later learning depends upon.
If technology bypasses these learning processes, children may produce answers without developing the underlying cognitive abilities required to understand them.
This concern is particularly relevant for generative AI. Unlike calculators or spell-checkers, AI can complete entire tasks on behalf of students. It can write essays, solve maths problems, summarize texts, and generate explanations instantly. While this can be useful for mature learners, it also creates a temptation to skip the productive struggle through which learning actually occurs.
Educational psychologist Daniel Willingham has frequently argued that thinking depends heavily on knowledge already stored in long-term memory. Students cannot develop critical thinking skills in a vacuum; they need a substantial foundation of knowledge first.
In other words, before children can effectively use AI as a tool, they need enough understanding to evaluate whether the AI is correct.
Why Handwriting Still Matters
One of the most overlooked casualties of the digital classroom has been handwriting.
Research consistently shows that handwriting engages the brain differently from typing. Writing by hand requires students to process information more actively, strengthening memory formation and conceptual understanding.
Studies have found that students who take notes by hand tend to retain and understand information more deeply than those who type verbatim notes on laptops. The slower pace of handwriting forces learners to summarize, process, and encode information rather than simply transcribe it.
Handwriting is not merely an old-fashioned skill. It is part of how the developing brain learns to organise language, attention, memory, and motor coordination.
When young children spend less time writing by hand, they may be missing opportunities to strengthen these foundational neural pathways.
Paper Still Has Advantages Over Screens
The same pattern appears in reading research.
A major meta-analysis published in Computers & Education found that reading on paper generally produces better reading comprehension than reading on screens.
Research specifically involving children found that reading from screens impaired comprehension compared with reading from paper, and notably, children were often unaware that their understanding had declined.
More recent studies have continued to show that screen-based reading often encourages shallower processing and skimming behaviours, while paper reading promotes deeper engagement with the text.
This does not mean screens are inherently harmful. Digital tools can be extraordinarily useful. But the evidence suggests that the medium itself influences how deeply children process information.
When literacy skills are still developing, those differences matter.
Are Children Becoming the Subjects of a Giant Experiment?
Perhaps the most important lesson from Norway, Finland, and Sweden is not about AI specifically.
It is about caution.
Educational technology is often introduced at scale long before we fully understand its long-term effects. New devices, platforms, apps, and now AI systems are frequently adopted because they are innovative, not because there is strong evidence that they improve learning outcomes.
The result is that entire generations of children can become participants in large-scale educational experiments.
Only years later do we begin asking whether these tools actually improved literacy, numeracy, attention, memory, or critical thinking.
By then, the experiment has already been conducted.
Back to Basics Doesn’t Mean Backward
Norway’s AI restrictions should not be interpreted as a rejection of technology.
In fact, the policy explicitly recognises that older students need to learn how to use AI responsibly. Students aged 17–19 will receive guidance on using AI appropriately for higher education and future employment.
The distinction is important.
Technology should build upon strong foundations, not replace them.
Children first need to learn how to read before AI reads for them. They need to learn how to write before AI writes for them. They need to learn how to solve problems before AI solves them for them.
Only once those foundations are secure can technology become a powerful amplifier of human capability rather than a substitute for it.
Perhaps Norway’s decision is a reminder of something education has always known but occasionally forgets:
The newest tool is not always the best teacher.
Sometimes progress means returning to the fundamentals that helped children learn successfully in the first place.
References:
- Reuters report on Norway’s AI restrictions in schools.
- Kong, Seo & Zhai (2018), Comparison of Reading Performance on Screen and on Paper: A Meta-Analysis.
- Halamish & Elbaz (2020), Children’s Reading Comprehension and Metacomprehension on Screen Versus on Paper.
- Clinton (2019), Reading from Paper Compared to Screens: A Systematic Review and Meta-Analysis.
- Jensen, Roe & Blikstad-Balas (2024), research on screen reading and shallow processing.