BeLearn Take-Home Exams
Examinations in the age of ChatGPT: this project examines how generative AI can be integrated responsibly into competency assessments.
Factsheet
- Schools involved School of Engineering and Computer Science
- Institute(s) Institute for Patient-centered Digital Health (PCDH)
- Research unit(s) PCDH / AI for Health
- Funding organisation Others
- Duration (planned) 20.01.2025 - 31.01.2026
- Head of project Prof. Dr. Kerstin Denecke
-
Partner
BeLEARN
Universität Bern - Keywords Künstliche Intelligenz, Large Language Model, Bildung
Situation
The project investigates how to responsibly integrate generative AI – in particular ChatGPT – into competency assessments at universities. It tackles the challenge of protecting academic integrity whilst embracing the educational potential of AI-powered tools. Ultimately, the project aims to design, trial and evaluate new take-home exam formats that ensure examination practices that are fair, transparent and conducive to learning.
Course of action
The project follows a multi-phase research design that combines literature review, development, piloting and evaluation. Based on existing research, two innovative examination formats are being developed using ChatGPT and tested in actual university courses. Qualitative and quantitative methods are used to examine their feasibility, effectiveness and impact on learning. In parallel, practical workshops will train lecturers in how to apply and adapt the formats. An iterative approach ensures that the results are continually refined further and remain relevant to practice.
Result
Based on a targeted review of the literature, we developed a comprehensive approach to the responsible use of generative AI in take-home examinations. This led to two main results: (1) An ILIAS self-study course on academic writing using generative AI (e.g. prompting, structuring, quality assurance), which incorporates ongoing documentation of the nature and extent of AI use, as well as a personal prompt portfolio as part of the assessed work; and (2) the design and piloting of an ‘AI for Students’ day as an entry-level event to help students build foundational skills. The self-study course was integrated into two different modules (master’s/bachelor’s; different formats). These were piloted and evaluated through pre- and post-course self-assessments as well as verbal feedback, confirming the expected increase in student competency while also identifying and resolving minor technical issues. The ‘AI for Students’ day received good feedback from the 120 or so who attended and will be continued at a later date.
