I was recently invited to the subject meeting in Business Administration to speak about AI in teaching and assessment. Colleagues wanted to hear more about the discussions taking place at management level, at the Swedish Higher Education Authority, and within the Ministry of Education.
They were curious about the expectations industry has of newly graduated students. They also wanted a concrete discussion on how the concept of “independent project” in the Higher Education Ordinance should be interpreted in light of AI, and to learn from experiences in other subjects and at other institutions. The meeting began with Niklas Jakobsson, Professor of Economics, explaining that he had generated a fully acceptable business administration thesis in roughly two hours with the help of AI. It was an excellent starting point for the discussion.
Our conversation came to revolve around quality assurance and the requirement for progression within our programmes, what employers actually expect from students now being educated with AI as an everyday tool, and practical issues such as exam halls and what we communicate on our website. These are conversations currently taking place across many subjects and at many universities.
I presented the proposal for a Policy for the Responsible Use of AI, which states that the institution should neither encourage uncritical use of AI nor distance itself from it. The position is that AI should be regarded as a tool that always requires professional judgement regarding appropriateness, consequences, and responsibility in each specific context. In some courses or components, it may be pedagogically or scientifically justified to work entirely without AI tools. That is a fully legitimate choice.
The issue that engaged people the most concerned degree projects and the interpretative space around the concept of “independent project” in the Higher Education Ordinance. What does independence mean when an AI tool can produce a coherent and well‑supported text in a matter of hours? Is it the process that must be independent, the product, or both? And how do we assess that? We are clearly in a period of transition which ultimately is not about AI itself but about what a degree project is meant to show, and what we actually want to assess.
The policy proposal is clear on this point: AI must not replace independent learning or scientific ability. But how this is put into practice in each subject is precisely the collegial dialogue we need to have. There are interesting examples to learn from, such as revised assessment formats and requirements for transparency regarding AI use. But there is no universal solution to copy.
These types of collegial conversations need to take place within every subject and at both local and national levels. They must be anchored in the contexts where teaching and assessment actually occur. I also encourage everyone to engage with the research that already exists on AI in higher education—what we know, and what we don’t. It provides a solid basis for discussing the future of learning and assessment. The policy contains nine recommendations and is now out for internal consultation. Contact your faculty office to find out how you can contribute.
