Björn Svensson, Director of studies at the Department of Informatics at LUSEM, has early on shown great interest in the opportunities and risks that generative AI may bring to university education. He is also a member of the LUSEM AI committee, whose purpose is to develop AI-related guidelines and policies and to give advice to the various departments at the faculty.
During the programme, Björn discussed how universities can counter AI cheating and highlighted the importance of integrating AI responsibly. He mentioned that Lund University is working on adjusting examination methods, for example by requiring students to orally answer questions in addition to written submissions. This is to ensure that students actually understand the answers to the assignments they have submitted.
When asked about the benefits of using AI at university, Björn replied:
“You can absolutely use language models like ChatGPT and Gemini and Anthropic's Claude to mislead and cheat, but they are also great tools to help and support the students' learning process, provided you are mindful of hallucinations and use them responsibly.“ (Translated from Swedish.)
After the broadcast, Björn thanked TV4 for the invitation and shared further thoughts on LinkedIn:
The key, I think, lies in creating incentives for responsible AI use in education, showing students good examples of use (prompting), and explaining and showing that individual knowledge is still important – especially when using language models!
(translated from Swedish)
