AI is fundamentally changing how we teach, conduct exams, and think about student formation. This week's news takes a close look at the tension between technological progress and the purpose of schooling.
This week we have compiled a range of news and articles on AI in teaching that together paint a picture of an education system in fundamental transformation. It spans from upper secondary school (gymnasium) leaders’ calls for AI pilots to teachers’ concern about being replaced.
One of the biggest challenges is an examination system that increasingly does not reflect students’ reality.
Last week was the Folkemøde (the People’s Meeting) on Bornholm, and unfortunately we were not able to attend. But the debate about AI’s role in teaching was very much discussed, and here are two articles that caught our eye.
New studies from both Denmark and abroad reveal a mismatch between grand AI visions and everyday challenges.
While the debate about AI often centres on practical challenges, more voices are now raising the question of the broader educational mission. Should digital understanding be a cornerstone in school, or do we risk losing something essential?
We also have this week’s news from the tech world. Finally, you will find the articles we did not have room to expand on.
Happy reading!
Maja Bødtcher-Hansen (Danske Gymnasier) and Niels Yde (Danske Erhvervsskoler og Gymnasier) are calling for more action and more concrete trials with generative AI in the upper secondary school (gymnasium). They argue that both teaching and especially assessment formats are lagging behind technological developments, and that AI has created an imbalance between the examination system and students’ actual ways of working.
Niels Yde and Maja Bødtcher-Hansen point out that the current culture of control and mistrust between teachers and students is counterproductive and hampers learning outcomes. Among the proposals is that AI should not only be used by students, but also employed as a tool by teachers and examiners in the assessment of written assignments.
They also recommend broader pilots with case-based exams, portfolio exams and combinations of written and oral components. Danske Gymnasier also highlights the need to make technology comprehension a standalone subject in the upper secondary school (gymnasium) to ensure students’ education and readiness for further study.
Text_ Johan Rasmussen Photo_ Danske Erhvervsskoler og -Gymnasier and Jacob Nielsen
At Støvring Gymnasium, upper secondary students have sat a written Spanish exam without using a computer, as part of a pilot scheme that from next school year will become permanent in the beginner languages Spanish, Italian and French. Here, teachers and students report increased motivation and renewed engagement in language teaching when everything is done with paper and pencil.
Although many students experienced physical discomfort when writing by hand and were unsure about their handwriting, they have generally been positive about the exam format. Teachers judge that students’ handwritten Spanish is generally legible, even if it sometimes requires extra effort. At the same time, according to Spanish teachers and the school’s headteacher, it has provided a more honest and down-to-earth learning experience.
Students with special examination arrangements still have access to digital aids.
Mads Søndergaard Thøgersen
During a debate at the Folkemøde, organised by Danmarks Lærerforening (the Danish Union of Teachers) and covered by DR, it became clear that views on the use of AI in the Danish primary and lower secondary school (folkeskole) are sharply divided. Some see AI as a major opportunity to strengthen teaching, while others fear that it could undermine students’ motivation and the very purpose of school.
Teacher Christel Rasmussen, for example, pointed out that AI can remove the point of learning subjects such as German, since the technology can already translate. She believes that AI necessitates a fundamental rethinking of how school should function.
Panellists such as Regitze Flannov and Alexander von Oettingen acknowledged that AI can strengthen students’ critical and creative competences, but they also pointed to the need for political steering and continuing professional development.
Maria Becher Trier
In a debate at the Folkemøde titled “Goodbye teacher, hello chatbot?” lawyer and AI instructor Brian Nygaard warned that teachers risk becoming redundant within five to ten years.
He emphasised that AI can not only convey knowledge, but also simulate emotions and understand human psychology at a level that makes the technology suitable for teaching, perhaps even in physical form as humanoid robots.
Language and maths teachers are seen as particularly vulnerable, because AI can translate and solve tasks more efficiently. Brian Nygaard also questioned whether it will even be necessary to learn mathematics and languages in the future, when AI can carry out the tasks.
His perspective met resistance from Jacob Gjelstrup from Uddannelsesforbundet (the Danish Union of Education Professionals), who stressed that the teacher’s role cannot be reduced to knowledge transfer. He pointed out that the human relationship and pedagogical presence are central elements of teaching that AI cannot replace.
The debate illustrates the widely differing views on AI’s role in teaching and raises fundamental questions about what tomorrow’s school should look like, and which parts of the teacher’s role are irreplaceable.
A nationwide survey of Norwegian teachers shows that the use of AI in writing instruction often takes place without clear guidelines or a shared practice. The survey, carried out by Skrivesenteret ved NTNU, covers 530 teachers from across Norway and shows that almost 80 percent of teachers have taken the initiative themselves to acquire knowledge about AI. Many teachers use tools such as ChatGPT, but almost half do not use AI in teaching at all, often due to a lack of local policy or support from leadership. The survey concludes that the absence of clear guidelines creates large variation in both use and attitudes.
Some teachers report outright bans on AI, while others experience pressure to adopt the technology.
Anne Holten Kvistad
A new investigation by The Guardian reveals that nearly 7,000 British university students were caught cheating with AI tools such as ChatGPT in the 2023/24 academic year, representing a marked increase compared to the year before.
That equates to 5.1 cases per 1,000 students and is expected to rise further in the coming academic year.
At the same time, the number of instances of traditional plagiarism has fallen markedly.
The investigation indicates that AI-enabled cheating is harder to detect than classic plagiarism, making it more difficult for universities to police.
Michael Goodier
The chair of Skolelederforeningen (the Danish School Leaders' Association), Dorte Andreas, argues in an opinion piece that digital literacy should have the same status in the Danish primary and lower secondary school (folkeskole) as reading and arithmetic. She points out that AI is already a natural part of both the classroom and children’s leisure time, and that the school therefore has a particular responsibility to equip pupils to navigate a digital world with integrity and critical judgement.
The technology offers many possibilities, but also serious challenges, especially when it distorts reality through algorithms and AI-generated images and answers. She cautions against leaving it to individual teachers to integrate digital literacy into an already busy workload, and calls for a national strategy with investment and political will.
Teachers need support through professional development and clear frameworks, so that technological understanding is neither ad hoc nor deprioritised.
School must not become a place where tech giants experiment on children, but rather a protective environment where pupils learn to understand technology and its impact.
Professor Morten Goodwin from the University of Agder in Norway argues that, in future, pupils should be able to understand and use AI on a par with reading and mathematics.
He criticises the tendency for schools to try to shield themselves from the technology, warning that this merely leaves the field to global, commercial actors with no pedagogical grounding.
AI should not be kept out of school, but actively and meaningfully integrated as part of pupils’ learning journey. In a digital reality marked by deepfakes, algorithmic influence and AI-generated content, it is necessary to strengthen pupils’ digital judgement and critical thinking.
That must happen in the classroom, not on social media or through unregulated apps.
Morten Goodwin
Victoria Livingstone and Jeppe Klitgaard Stricker warn in a new article at UNESCO that AI removes the “cognitive friction” – the slow and often frustrating process through which students learn to think critically.
They use a concrete example in which a student spent a whole month investigating the “Mozart effect” and ended up with a nuanced conclusion about how debunked research can still have value. With AI she could have had the answer within minutes, but would have missed the entire learning process.
The authors see a worrying trend: when students learn to ask “AI-friendly” questions to get good answers, they risk dropping the complex and unclear questions that, in reality, lead to deeper learning.
The problem is not AI itself, but that we start valuing questions by how well they work with the technology – rather than by intellectual curiosity.
"As educators, we must ensure that the purpose of questioning does not collapse into the mechanics of prompting. It is pivotal that students continue to value questions, not for their compatibility with AI, but for their capacity to open up new ways of thinking."
Their solution is a two-part model, where students first use AI to get an overview and frame topics, and then work with real sources, discuss and reflect.
A new study from MIT shows how repeated use of GPT-4 for assignment writing can negatively affect brain activity and cognitive abilities. In the experiment, 54 students wrote SAT-style essays across four different sessions while their brain activity was monitored with high-resolution EEG.
Participants were divided into groups that used either no help, Google or GPT-4. In the fourth session, the groups switched tools to assess the effect of sequence. The results show that using GPT-4 alone reduces the brain’s network activity, especially in areas related to creativity and language comprehension.
Students who first wrote on their own and later used AI, by contrast, showed increased brain activity and greater creative variation. Conversely, those who started with AI and later had to write without help showed lower coordination and less linguistic variation, even though their essays were rated highly.
The study concludes that repeated reliance on AI creates a kind of “cognitive debt” – a temporary gain in efficiency at the expense of the ability to think deeply, remember, and be creative in the longer term.
According to the researchers, sequence matters: starting with your own ideas and only then using AI as support can preserve the brain’s creativity and critical thinking.
Nataliya Kos’myna
Here is a brief overview of new or improved AI technologies that emerged last week.
Ritzau
Peter Christian Bech-Nielsen
Theron Mohamed
Maximilian Schreiner
Department for Education
