DF's Alex Ahrendtsen proposes using AI in academic advising on the new EPX program. We point out, however, that AI-based advising is classified as high-risk by the EU and cannot replace the human relationship in academic advising.
In the United Kingdom, in 2020 an algorithm was implemented to predict students’ exam grades during the COVID-19 pandemic. The result was a national scandal, as the system systematically downgraded students from disadvantaged areas. Thousands of young people lost their university places until public anger forced the government to withdraw the algorithm. This is just one example of how automated decision-making systems in the education sector can have serious consequences.
With this experience in mind, we should be particularly attentive when Dansk Folkepartis undervisningsordfører (education spokesperson), Alex Ahrendtsen, in Gymnasieskolen proposes implementing AI in guidance counselling on the new epx-uddannelse. Ahrendtsen states that “artificial intelligence can be a good supplement to guidance counselling”, comparing it with websites such as Uddannelsesguiden (the Danish education guide). He adds that AI “might also give guidance counsellors more time to talk with students”.
These statements leave several unanswered questions: Are we talking about simple search tools with AI enhancements – or advanced systems with personalised recommendations for educational choices? The lack of detail makes it difficult to assess the real implications of the proposal.
At the same time, Ahrendtsen’s own remarks suggest that financial considerations play a significant role: “The government is not in favour of having more guidance counsellors in education – that costs money.” He continues: “There were so many other things that were more important to us in the negotiations, and so I didn’t want to go down that road. That’s why I put forward the idea of AI as a supplement.”
This link between the AI proposal and financial considerations raises the question of whether the technology is seen primarily as a cost-cutting exercise rather than a well-thought-out pedagogical solution. It is hardly a coincidence that Ahrendtsen wants AI guidance introduced in the new epx-uddannelse, which will not open until 2030. This creates a situation in which the trial cannot be compared with existing practice and experience. Should young people in Denmark be guinea pigs for untested technology? And is it really about creating a model that can later be rolled out across the entire education system – with savings as the real success criterion?
Whatever specific AI solution Ahrendtsen has in mind, it is crucial to consider the EU’s AI Regulation. The Regulation clearly categorises AI systems used to assess or provide guidance on educational choices as high-risk systems.
Such AI systems are subject to a number of specific requirements:
Labelling an AI solution as a “supplement” does not exempt the system from these requirements if it provides recommendations about educational choices.
Even the most advanced technology cannot replace the personal element in guidance. As Kåre Blinkenberg puts it: “Guidance is about relationships between people. It is about the student’s motivation, intentions and dreams, and we cannot outsource that to a machine.”
Lotte Klein from Ungdomsuddannelsernes Vejlederforening (the Association of Guidance Counsellors for Youth Education) likewise emphasises the need “to ensure professionalised guidance for students in education”.
Young people with complex backgrounds, diagnoses, or ordinary uncertainty about their educational choices are particularly in need of the empathetic dimension in guidance – to be seen, heard and met as a human being. An AI-based solution cannot pick up the nuances of body language, tone of voice, or the unspoken concerns that often lie behind an educational choice.
In our assessment, there are two realistic ways to use AI in guidance counselling.
If it is merely an improved version of Uddannelsesguiden (the Danish education guide), then it should be called what it is, namely an information tool – not an AI guidance system. It may be valuable, but it is not revolutionary and is unlikely to free up time for counsellors.
But if Ahrendtsen actually envisages a system that will provide personalised recommendations for education, then he must address the AI Regulation’s requirements for high-risk systems. That entails extensive data security, human oversight, transparency in the algorithms, regular audits, and substantial funding.
In both cases, we must insist that qualified human guidance counsellors remain central to the process. AI can be a supplement, yes, but not a cheap substitute.
Ahrendtsen’s proposal for AI in guidance counselling on epx is emblematic of a broader trend in the political debate that views AI as a magical solution to complex problems. But the EU’s AI Regulation has, for good reason, placed educational guidance in the high-risk category.
The Regulation states clearly: “AI systems used within education or vocational training, in particular to determine access or admission… can be particularly intrusive and violate the right to education as well as the right not to be discriminated against, and make historical patterns of discrimination permanent.” (Source: Official Journal of the European Union).
This framing should be the starting point for any serious discussion of AI in education. It should not be seen as an obstacle to innovation but as the basis for the responsible implementation of new technology in an area that is of decisive importance for young people’s futures.
This article was published in Gymnasieskolen:
Per Størup Lauridsen
Text: Johan Rasmussen Photo: Shutterstock
Richard Adams
Matt Burgess
