Universitetsavisen
Nørregade 10
1165 København K
Tlf: 35 32 28 98 (mon-thurs)
E-mail: uni-avis@adm.ku.dk
—
Education
Generative AI — Chatbots can give us quick answers. But the university should protect the slow and difficult process of arriving at them, writes a master’s student in Education.
Last week, University Post reported that the management of the University of Copenhagen has decided to invest DKK 110 million in a major AI initiative, headed by the recently appointed Pro-Vice-Chancellor for AI, Morten Axel Pedersen. Over a three-year period, he will work to integrate generative AI into research, education and administration at the university.
Opinion
This is an opinion piece. It reflects the author’s own views.
We encourage everyone to read opinion pieces in full before commenting on them on Facebook, so that the debate remains constructive.
Disagreement is welcome, but please keep the tone respectful.
University Post reserves the right to delete comments that violate our debate guidelines.
It is striking that UCPH can find DKK 110 million for the project when, in recent years, the university has implemented major budget cuts amounting to several hundred million kroner.
The finances, however, are the least concerning aspect. The interview with the new Pro-Vice-Chancellor gives the impression that UCPH’s AI initiative is based not on the existing knowledge and research in the field, but on assumptions, factually incorrect claims and false hopes about what generative AI can be used for.
In the interview with University Post, Morten Axel Pedersen says: »I don’t think it will turn out that we lose brain cells by using AI. And I doubt that it will make us fundamentally less intelligent.«
We may not lose brain cells by using generative AI, but research suggests that it has a negative impact on our ability to think both creatively and critically.
It is also clear that the university management is failing to engage with current research highlighting the negative impact of generative AI on learning and learning communities, including the growing distrust between students and teaching staff.
At the Faculty of Humanities, where I study, the guidelines on the use of generative AI state that the technology may be used for »dialogue and feedback«, or to answer questions that students would otherwise ask their fellow students or teachers.
But dialogue takes place between people, not between a person and a technology. Something unique emerges in dialogue with another person because we all bring our own histories and prior understandings to the conversation.
Why should I ask a chatbot the questions I have instead of asking my fellow students or my teacher?
Why is UCPH promoting an isolating and individualistic culture among students in this way, instead of encouraging them to talk and discuss with one another?
We live at a time when more and more people feel lonely, and we need more than ever to talk to each other — not to a chatbot.
According to Morten Axel Pedersen, it is time for UCPH to engage seriously with generative AI because »most people who graduate with a master’s degree will, to some extent, become a kind of manager of AI agents.«
Is this really the University of Copenhagen’s view of what its graduates should be capable of? That we should become a kind of robot tasked with managing a technology? It sounds incredibly dystopian.
The university’s primary purpose is not simply to produce graduates for the labour market. Universities should be critical institutions that produce knowledge.
The idea that generative AI is inevitable and that universities therefore have no choice but to keep up is based on a false premise. It is the major tech companies that are selling the idea of inevitability, and generative AI is only inevitable if we accept that premise.
UCPH’s AI initiative promotes a view of learning and education that is best compared to old-fashioned rote learning: The »all-knowing teacher« is simply replaced by an »all-knowing technology« from which we students are expected to passively receive and reproduce knowledge.
We know that increased use of generative AI will lead to less immersion and critical thinking. Generative AI reproduces hegemonic narratives, not curiosity and critical thought.
Yes, it is easy to get an AI to answer a question. But learning is much more than the final conclusion. In our eagerness to embrace new technology, we risk skipping crucial steps: reading the material, discussing it with fellow students, making mistakes, encountering counterarguments and arriving at a different conclusion from the one that initially seemed obvious.
Learning is the long and complex process of reaching a conclusion. It takes time and can be difficult, but we do not learn to think and reflect critically if we outsource the intermediate steps to a technology.
Learning is not simply the transfer of knowledge from one actor to another. It also arises through encounters with fellow students and teachers — when we ask questions, discuss, misunderstand and are challenged.
In its enthusiasm for generative AI, UCPH risks streamlining away the very thing the university should protect: thinking together.