Abstract:

Bifröst University in Iceland is facing challenges due to technological developments, including artificial intelligence. There have been significant changes in the last few years. Lectures and teaching are asynchronous, and students access lecture recordings and other course materials via the Canvas LMS, which the University implemented during the 2019-2020 academic year.

Data was retrieved from the LMS for the academic years 2020-2021, 2022-2023, and 2024-2025 and used to conduct a comparative analysis of syllabi. Using the datasets, the study investigates whether assessment patterns indicate a shift away from traditional final examinations toward more varied forms of continuous and applied assessment. The analysis is based on course-level assessment descriptions published on the syllabi pages. These were systematically extracted and coded for the presence of assignments, written exams, oral exams, final projects, participation-based assessment, presentations, pass/fail grading, and explicit statements that no final exam was used.

The findings show that coursework-based assessment is nearly universal across the sample, while exams remain common but no longer function as the sole or even dominant basis of assessment in many courses. Approximately 68% of courses mentioned some form of examination, 51% included a written exam, 26% included a final project, and 30% incorporated participation, discussion, or attendance-based components. Between 2021 and 2025, the relative presence of presentations, oral assessment, and project-based components increased. Comparison of repeated course codes across years suggests that this change cannot be explained solely by differences in course mix. Several courses appear to have shifted from written or oral final exams toward assignment-based, collaborative, or staged assessment structures. The most notable development is expansion of blended assessment designs that distribute evaluation across multiple activities, including projects, discussions, and iterative coursework. These findings contribute to current debates on assessment reform in higher education by offering a longitudinal data collection.

This is a good start for further analysis of the development of assessment practices post AI in higher education. In addition, the University’s business model has changed: since the 2024-2025 academic year, students no longer pay tuition fees; instead, they pay a significantly lower registration fee annually, resulting in exponential growth in the number of students. The authors are interested in seeing what effect these developments, if any, have had on the development of assessment practices at the University. The next step would be to check possible patterns based on course size and/or different disciplines. Another comparison would be graduate vs postgraduate.

Keywords:

Assessment, higher education, AI.