Artificial intelligence in school development – the Netherlands as a role model

This article is a translation from a German article into English, with assistance of the AI based translator Deepl. Here is the original Article: https://www.campus-schulmanagement.de/magazin/kuenstliche-intelligenz-im-unterricht-das-vorbild-niederlande

Prof Inge Molenaar reports on how far the topic of artificial intelligence in the classroom has progressed in the Netherlands.

How can artificial intelligence (AI) be integrated into everyday teaching in schools? The Netherlands is already several steps ahead of Germany in this respect. Prof. Inge Molenaar from Radboud University Nijmegen explains in an interview how the neighbouring country has achieved this and what the next steps are.

Editorial team: Professor Molenaar, how far has the integration of artificial intelligence progressed in schools in the Netherlands compared to Germany?

Prof Inge Molenaar: I think there is a relatively big difference between Germany and the Netherlands when it comes to the availability of appropriate technology in schools. Pupils in the Netherlands work more frequently with digital devices, most children in primary schools use tablets, pupils in secondary schools have their own digital devices. Almost every form of teaching and learning material is already digitalised, which has also been very helpful during the coronavirus pandemic.

This gives us a very good starting point for driving forward the implementation of artificial intelligence in schools and puts us in a leading position in Europe. We are already making extensive use of adaptive learning technologies – ALT for short – in primary schools. They adapt the learning material to the needs of the learners and give direct feedback on whether something is right or wrong. ITS (intelligent tutorial systems, editor’s note), on the other hand, are still rarely or not at all in use. In contrast to ALT, ITS personalise on several levels, adapt the material, provide detailed feedback and enable fully personalised learning processes.

Editorial team: The fact that Dutch schools are so far ahead did not happen overnight. In your presentation at the AI conference in Berlin, you mentioned that the Netherlands has been working on the digitisation of schools on a large scale since 2014.

Molenaar: That is correct. What is important here is that this process was scientifically accompanied from the very beginning and the corresponding technology was repeatedly adapted and optimised in constructive dialogue with practitioners at the schools. For example, questions were asked about how the schools use the technology and programmes, why certain options are not used and which programmes or functions cause difficulties. It was therefore about technical changes, but also about didactic improvements. It is often assumed that bringing technology into schools is enough. The opposite is true – because the work only begins with the technology on site: the learning, the development, the fine-tuning, the innovation.

Editorial team: What was the public reaction in the Netherlands to the latest AI developments such as ChatGPT? In Germany, schools were initially quite critical of the topic.

Molenaar: We initially had a similar reaction. Mainly because many realised that students could now have their homework done by AI. So there was this realisation that the previously used, carefully developed teaching and learning processes would be largely undermined by artificial intelligence such as ChatGPT.

As a result, some schools tried to ban these programmes before they understood that they wouldn’t work. So many schools began to deal with the issue differently and ask themselves questions such as: How can homework be modified so that it cannot be generated by AI, or can only be generated poorly – for example, by asking personalised questions or making certain connections that AI cannot easily solve. At the same time, the topic was integrated into teaching and discussed with the students: How should one work with programmes such as ChatGPT? What can such a programme do and what can’t it do? Pupils also asked themselves: I will be working with such systems for the rest of my life. How can I learn how to use them if my teachers don’t teach me?

Editorial team: What risks and challenges do you see when it comes to AI in schools?

Molenaar: Of course we are thinking about what AI in schools means for the privacy and safety of pupils. A lot of the data goes directly to the USA via servers. So can we develop buffers so that we can protect data and keep it here in Europe and still have these AI tools available to us? At the same time, the issue of the risks of this technology in the education system is also about autonomy, i.e. the question: How do we ensure that AI systems do not restrict individuality and humanity?

How do we prevent pupils from becoming robots in their own learning? How can we ensure that teachers stay up to date and understand how AI works and how they can use it to design hybrid lessons? These topics are addressed through professionalisation programmes and dialogue between science, business and educational practice as part of the National Education Lab AI in the Netherlands.

Editorial team: You already mentioned that the Netherlands is leading Europe when it comes to AI in schools. Where do you see Europe in a global comparison with regard to the integration of AI in schools?

Molenaar: The big problem we have in Europe is that we have a very diverse, fragmented educational landscape. In Germany alone, there are 16 federal states with their own requirements, as well as all the other countries with different languages, curricula and markets. That is why it is very important for the National Education Lab AI that we develop a European home market so that our developments and progress in the field of AI can grow sustainably here. Because we need these own developments.

During my visit to China, I realised how quickly the development of AI is progressing there. At the same time, its use there is not the kind of AI and not the kind of teaching methodology that we want in Europe – but it is one that is being strongly developed and has already found its way into European schools equipped with Chinese safety technology. That must give us cause for concern. And there are also developments coming from the USA that do not correspond to our values and our system. It is therefore important that we in Europe make progress, maintain control and go our own way.

Editorial team: How do you see the role of the teacher in a classroom characterised by artificial intelligence?

Molenaar: Who says that teaching has to be characterised by AI? For me, the role of the teacher is very important. As in Germany, teachers in the Netherlands are also under a lot of pressure due to the shortage of teaching staff. While teachers were initially very keen to stay in control of AI, I now hear more and more often that teachers see AI as a helpful support to counter the shortage of teachers.

I’m not a big fan of this view because it takes teachers to ensure that AI is used in the right way in schools. It’s about finding the right balance between teacher-led lessons and AI input. To do this, we need an ecosystem where there can be a dialogue about how AI is used in the classroom – from the teacher’s perspective. When teachers are scarce and it’s all about achieving the best results, the system and the balance needed for good teaching with AI come under pressure.

Editorial team: What positive developments and opportunities does the topic of AI bring to schools? And what are the next steps that need to be taken in dealing with AI?

Molenaar: There are many exciting and creative examples of how AI can enrich the classroom. For example, through dialogue writing with an AI or exploring the different plants in the natural environment with an AI that tells you the type of plant, the family and the relationships between them. But to take teachers there, it is important to meet them where they are today.

We are very satisfied with the level of digitalisation in the Netherlands, now it’s about making the computer programmes smarter. This includes more sensitive, more detailed feedback from the systems, the possibility of customisable tasks, the expansion of tools at secondary schools, where the requirements for AI technologies are becoming more complex and also require more research. These are all steps that we will be taking in the next two to three years.

If you zoom out, you also have to say that we are only good at one thing so far: understanding how a child develops knowledge. But focussing on cognitive development is not enough; this is only one part of education. Social and emotional development is also part of it, as is self-regulation. Not all important elements of learning can be captured in ones and zeros: Certainly not all the intuition and all the knowledge of teachers can be integrated into AI systems or even replaced. We still have many steps ahead of us. It is necessary to conduct research together with schools, teachers and scientists in order to make progress in this area.

Editorial team: Professor Molenaar, thank you for this interview.


About the person

Inge Molenaar is Professor of Behavioural Science Institute and Orthopedagogics: Learning and Development at Radboud University Nijmegen in the Netherlands. She is particularly interested in the interface between education and technology and how individualised systems can be developed to meet the specific needs of students.

And here is, as so often, a different mindset between The Netherlands and Germany: https://international.eco.de/presse/eco-survey-great-scepticism-towards-the-use-of-ai-in-german-schools/

Our modern and future focussed, is in Germany often a challenge, during interviews and communications.