AI Challenges Learning – and the Way We Teach

Dear staff

Last week, I helped organize a theme meeting in the Danish Academy of Technical Sciences titled What is AI doing to our educational programs? It is a topic that I sense many are affected by, and one that is frequently discussed in our teaching and education committees and, most recently, in the Academic Council – and which is part of our new strategy.

For the meeting, we brought together both practitioners and chairs from government committees, and the participants came from universities, university colleges, upper-secondary schools, and foundations. The meeting was well attended with nearly 100 participants. It resulted in some interesting cross-disciplinary discussions.

From the teachers’ perspective, the clear challenge is finding teaching methods that ensure students actually achieve learning. AI removes the effort associated with learning, as AI bots can now answer almost any question, and many teachers experience that written assignments are getting poorer, and more students fail oral exams.

Several speakers concluded that it is precisely the effort – or the “friction” – that is crucial for learning: the fact that you must find the right answers yourself, that you must reason your way to a conclusion rather than simply repeat your AI‑generated reply. However, it is far more difficult to answer how you actually create friction in teaching.

There was agreement that students must have the basic knowledge firmly in place before using AI, but once that foundation is established, AI can help improve assignment quality. So how does one determine what constitutes the basic knowledge that students need to have at their fingertips?

Here, Per Størup, Head of IT and Communications at Tietgen School, introduced a new concept – at least new to me. He suggests that each teacher must define what is “the indispensable” content of their subject. That is what students must learn. In this way, one can maintain a focus on subject knowledge while the technology remains a tool.

Another challenge is keeping up with the rapid development of new AI bots. The AI‑bot market changes from month to month, and the tools are getting better and better. Per Størup suggested involving students in the development of teaching methods, as they are often fully up to date on the newest bots.

It is, however, not always pleasant to acknowledge that, as a teacher, you are no longer the one who knows the most about the technology. Therefore, there may be a need for teachers to meet across disciplines to share experiences with new technologies and methods for achieving good learning outcomes.

Here, I can appropriately recommend participating in the Science Education Hub, which frequently has AI on the agenda.

There are many concerns about learning in the context of the rapid growth in AI use within the education system, but as Anders Malthe‑Sørensen from the advisory committee to the Norwegian government put it so beautifully: students must still achieve the same as always – genuine learning is about understanding and thinking, not about reciting material.

Marianne Holmer, dean