AI is challenging our fundamental models of education and teaching

Recently, I have started receiving lengthy, well-structured emails with impeccable punctuation from senders who, just a few years ago, used to send me short, often messy emails with almost non-existing punctuation. These emails are easy to read; however, they take a long time to read and the content is usually rather generic and without any original insights. 

I am fully aware of the underlying reason for this change in style. Generative AI / GenAI. Text is being produced at an unprecedented rate, and this development has both pros and cons. 

Like all other universities, SDU is facing challenges in relation to AI in three areas: Administration, research and education. 

I am quite confident that we will manage to adapt to AI with regard to the areas of administration and research. It may not be easy, but in my view, we have reason to hope that we will succeed in reaping the benefits of GenAI without stumbling into too many pitfalls.  

As far as administration is concerned, I have no doubts that the well-oiled SDU machine will be able to reap the benefits of the efficiency and quality gains that artificial intelligence has to offer. As to research, the competitive pressure is so intense that new AI-based tools will almost automatically be adopted if they have the potential to enhance the research output and impact.  

To reiterate: it may not be easy – and we will certainly have to invest in encouraging the ‘proper use’ of AI – but in my opinion, administration and research will be the two ‘easier areas’.  

The most challenging area will be education and teaching. The main challenge is that we have to educate graduates who will be able to use AI effectively but at the same time will be capable of independent and critical thinking and of drawing on their deep understanding of the subject area they have studied. 

This means that we have to fundamentally rethink our methods of teaching and assessment. On the one hand, we must ensure that our graduates have been taught how to master AI. GenAI – and other forms of artificial intelligence – are essentially tools, and public and private sector employers rightly expect us to educate graduates who can translate the technology into tangible value.  

On the other hand, we must make sure that our students do not fail to acquire in-depth knowledge within their field of study because they had an AI write every single take-home assignment throughout their studies. Our employers have a reasonable expectation that our graduates have actually acquired knowledge and are capable of independent and critical thinking and on drawing on a deep understanding of the subject. 

The immediate – but often costly – solution is to convert exams to oral or written on-campus exams in what is known as ‘secure environments’. But this will hardly be enough. A lawyer, for instance, is expected to be able to think critically and to reason and argue effectively in relation to a specific case. However, any law firm that will employ them will expect this to be done in conjunction with relevant AI tools.   

We have only seen the first glimpse of the challenges AI has in store for us in the area of education. And the only people you absolutely cannot trust are those who claim they know exactly what is going to happen and what the solution is. 

Once the summer holidays are over, we at SDU will be taking the first steps towards formulating a ‘sub-strategy for AI’. The focus will be on the three areas.  

When I read long, well-structured and generic texts these days, I am reminded of what our core mission as a university actually is: to educate people who are able to express themselves correctly – and who can also engage in think critical, reflective and independent thinking. These skills will become ever more important in a world with AI.