Teachers in higher education are dealing with increasingly complex challenges in evaluating student contribution in the fast evolving digital world. Students have had access to abundant information since the AI revolution started. The aim of the lecture is to test different methods at Bifröst University in Iceland, utilising oral exams and podcasts. The objectives are to check if these methods can deal better with the AI assistance that students frequently use to help them write essays and answer traditional project questions.

Bifröst university is built on a digital ecosystem that heavily utilises online learning methods. The authors are both professors and have used action research methods to self reflect on their experiments. The focus is on connecting theoretical knowledge with its application to practice. At this university, oral exams are used individually or for group projects, where the group is given 30-60 minutes to answer intensive questions related to the project. The authors have, on the other hand, used podcasts in their teaching with pre-recorded audio and/or ´YouTube´-like podcasts with both audio and visual appearance. In a 6-week modular teaching program, two teachers reflect on essential themes from the weekly lectures in approximately 20-30 minutes, engaging in a critical dialogue. In the same university, some courses have resulted in students’ project work being developed into a podcast created by 3-4 students. The topics are either chosen by the participants or the teachers. A typical student method is to have at least a loosely knitted manuscript or a list of questions ready for the podcast discussion.

The main conclusions of these experiments are positive. The communication during oral exams is, in many instances, fruitful and can contribute to the learning process, helping to prepare students for participation in real-life situations. The communication between students during the podcast is, in many instances, fruitful and can contribute to the learning process, helping to prepare students for participation in real-life situations or increase the depth of their knowledge. This is partly a peer-to-peer learning method. The evaluation and assessment is valid from this experience. But the individual oral exams in these cases appear to be more suitable for smaller classes with fewer than about 35 students.

Based on these results so far it seems logical to utilise multidimensional assessment tools and incorporate diversity in examination methods. These trials show the useful development of classic educational tools. Oral exams are proven traditional tools, used mostly in language classes. Oral exams have negative aspects, such as the stress factor and the subjective nature of the assessment. The positive factors are instant feedback and deeper dialog with the student. In larger classes, the group oral exam can be a valuable tool to use this method in broader settings. Podcasts haven´t traditionally been used much in higher education. A podcast used in university project work can be viewed as training for a work-related environment where students need to present clear and valid arguments. This is critical in online environments where AI help in exams can threaten academic rigour and integrity. These experiments look like robust and diverse assessment methods. At the same time these methods can possibly train students in critical thinking and practical connection of knowledge to work related assignments.

Keywords: Oral exam, podcast, higher education, AI, practical application.