Aarhus University Seal

New recommendations on the use of GAI at AU

In the spring of 2023, AU's Committee on Education set up a working group to examine the challenges and opportunities associated with the use of generative artificial intelligence in educational contexts. A number of interesting recommendations have come out of this, which you can read about here.

AU’s Committee on Education has examined the challenges and opportunities associated with the use of generative artificial intelligence in educational contexts. Photo: AU Foto

The GAI working group:

  • Niels Lehmann, vice-dean, Arts (chair)
  • Carsten Bergenholtz, associate professor at Aarhus BSS
  • Mette Vinther Skriver, associate professor at Health
  • Niels Lauritzen, associate professor at Natural Sciences
  • Poul Ejnar Rovsing, associate professor at Technical Sciences
  • Kim Kusk Mortensen, Head of Studies Administration, Nat-Tech
  • Sondre Strømsheim Pauli, lawyer at AU Student Administration and Services
  • Tina Bering Keiding, Division Manager at the CED
  • Dorte Sidelmann Rossen, special consultant at the CED (administrative support)

The recommendations can be found in Danish in a 15-page report with a four-page appendix. You can read the report in Danish here.

The report establishes that generative artificial intelligence (GAI) is a largely unavoidable technology that is already being used in many workplaces. Therefore, it is important that all students at AU gain a fundamental understanding of the opportunities and limitations of the technology:

“Rather than rejecting the use of the new technologies, it is therefore important to find sound and fruitful ways to incorporate the use of GAI,” (p. 1) the working group concludes, among other things, in the report.

On this basis, the working group has proposed that AU change gears and move from a prohibitive approach to an inclusive approach.

Subsequently, the working group recommends gradually introducing the use of GAI in the programmes in two steps. First by adjusting the design of the exams within the framework of the existing academic regulations. Then by developing learning outcomes, teaching formats, and types of examinations that take into account the advent of GAI.

The educational circumstances

The report outlines the educational opportunities offered by GAI and encourages the programmes at AU to include “GAI literacy” in the graduates’ qualification profile. In addition, the report proposes how GAI could potentially be included in teaching, and it highlights the potential of GAI to help students in need of special educational support.

However, there are also challenges associated with GAI. For example, one might fear that some students will cut corners or not use GAI appropriately because they do not understand the limitations of the technology. Again, the working group points to the importance of “GAI literacy,” including being critical of the outputs that a GAI produces:

“Finally, it is also worth bearing in mind that the quality of GAI’s response largely depends on the quality of the ‘prompts’ and information it is fed. Mostly, an unsophisticated use of GAI will not lead to an adequate response to a well-developed exam question.” (p. 7)

At the same time, the report points out that the use of GAI may require reflections on the balance between knowledge-reproducing and knowledge-generating learning outcomes in the programmes. For example, there may be a need for an increased focus on learning outcomes that address the students’ ability to use GAI competently.

The exam conditions

Concerning exams, the working group has eight points in the report (pp. 7-10). They acknowledge the fear that GAI could increase the potential for cheating but also point out that: 

“The potential for cheating has increased considerably, but principally there is nothing different going on from already known methods of cheating.” (p. 8)

In addition, the report has a number of recommendations, including that AU adjusts some types of examinations and, eventually, reassess and further develop the most GAI-sensitive ones. However, the report points out that one must weigh the balance:

“[I]t goes against the fundamental didactic effort to incorporate GAI competencies into the programmes to search for types of examination that counter the use of GAI. Therefore, it is more a question of prohibiting the use of GAI in certain examinations, making some of the exams GAI resistant with small adjustments while allowing other examinations to include the use of GAI." (p. 9)

For this, the working group recommends the development of a guideline to assess which exams can and cannot tolerate the use of AI. A tripartition can be used as the guiding principle: “prohibit,” i.e. a total ban of the use of GAI, “allow with attribution,” i.e. permission for use within specified limits, and “encourage,” where use is promoted.

Finally, it is also the working group’s recommendation that teachers talk with their students about the importance of academic integrity and judicious use of GAI. A great deal of work remains to be done where also all teaching staff at AU must gain more insight into both the opportunities and limitations.

Want to learn more?

Among other things, the Committee on Education has, based on the discussions of the working group’s report, decided to relax the rules on the use of GAI for Bachelor’s projects, theses, final exams on continuing and further education programmes, and exams for which it explicitly states in the course description that students may use generative artificial intelligence. Read more about this here.

If you are teaching and would like to learn more about the use of GAI and chatbots in your teaching, you can sign up for the CED’s workshop on generative AI.

You can also read more about GAI and chatbots on AU Educate.

Read the indstilling fra arbejdsgruppe om anvendelse af GAI på AU.