Motivation
Within the Public Administration programme, there is a strong emphasis on written assessment. With the rise of generative AI, the reliability of this form of assessment is declining, and it is becoming increasingly difficult to determine whether students have developed their own understanding. Oral assessment is seen by lecturers as a promising alternative, but it is difficult to scale and also has limitations such as potential biases, additional psychological pressure on students and variation in marking. Could a form of oral assessment contribute to the development of authentic, AI-resistant assessment methods?
Hypothesis
Bart is using his innovation voucher to design, test and evaluate various forms of oral assessment. In doing so, he is experimenting with, among other things, oral exams, interactive dialogues, mini-presentations with a defence, debates and peer interviews.
It is expected that oral assessment, provided it is carefully designed, can be a valuable complement to written assessment. Whereas written tests primarily measure factual knowledge, oral assessments offer greater insight into application, reasoning and communication skills, and are potentially more resistant to the influence of AI. This pilot therefore explores how oral assessment can be implemented sustainably in a way that is objective, scalable and workable.
Desired solution
The aim is to develop evidence-based assessment methods that can be used even with larger student cohorts, provide a reliable assessment of knowledge, and keep the workload for lecturers manageable. At the same time, we are exploring ways to enhance the objectivity of assessment within an AI-driven educational context.
In this way, Bart hopes to develop assessment methods that provide a better understanding of what students really understand and are capable of, even in the age of AI. Oral assessment offers opportunities for greater depth, dialogue and reflection.
Plan of action
Bart’s project consists of four phases. In the first phase, various forms of oral assessment are developed based on the literature and design choices, and submitted to colleagues for feedback. In the second phase, these forms are tested and evaluated with students in an experimental setting. In the third phase, selected forms of assessment are implemented in teaching and evaluated again. Finally, in the fourth phase, the results will be compiled and shared in a final report containing transferable models.
The pilot focuses on the Bachelor’s, Pre-Master’s and Master’s programmes in Public Administration, with a particular emphasis on larger teaching groups. The findings will be shared with other programmes, so that the assessment methods developed can also be applied there as a basis for future-proof assessment in an AI context.