Experiment: Allowing the use of AI in Final Take-Home Exams

Stefan Henningson, Professor from the Department of Digitalization has taken a bold step to explore the potential of Generative AI in student assignments. We had a conversation with Stefan who shared his insights on an experiment that allowed his graduate students to utilize AI tools in their final take-home exams.

Introducing AI into the classroom

In the first video, Stefan introduces the context of the experiment. Students in the elective course “Transforming IT Management for Digital Business”, were given two options for their final individual exam. They could either complete the assignment as they had in previous years, without any AI assistance, or they could embrace the use of generative AI, with the condition that they include a “transparency appendix” to explain how they used AI tools in their work.  

Despite some uncertainties about how their work would be evaluated, more than 50% of the students chose to experiment with AI.

In the video below, Stefan discusses the two options given to students for their final exam and the rationale behind introducing AI as a tool in the assessment process.

Reflecting on the experiment

So, how did the use of AI go? Did it enhance the students’ work, or did it fall short? In the following video, Stefan discusses some of the main findings based on the transparency appendixes where students explained how they incorporated AI into their assignments. He also shares an interesting observation about a subtle shift in their writing approach.