Effects of Generative AI on the Conception of the Course and the Assessment of Examination Results
If students are able to use AI in examinations, this has far-reaching implications. Not only the assessment of examination performance, but also the design of the course must be aligned with this, whereby various aspects must be taken into account:
Firstly, it must be ensured that fundamental principles such as equal opportunities and transparency are upheld (see here (external link, opens in a new window)).
Equal opportunities
Even with new technical possibilities, the principles of fairness and equal opportunities continue to apply. For example, the use of a (better) AI tool must not have a significant influence on the examination performance to be achieved. The examination must therefore be designed in such a way that comparable examination conditions exist for all students.
Transparency
The course must provide information in good time as to whether and for what purposes the use of AI is permitted as an aid in the examination and how its use is to be documented. We recommend using the University of Bamberg's AI policy generator (external link, opens in a new window) to create course-specific regulations. In addition, the purpose of rules and prohibitions as well as the possible consequences of violations should be explained. These can range from deduction of points or failure of the examination to refusal of further examination attempts and thus de-registration in the event of repeated or serious cheating.
In addition, the proof of the competences acquired and thus the assessment of the students' own performance is of central importance. Against this background, the question of which learning objectives and competences should be at the centre of the course must be clarified when designing the course.
It is important to check whether the use of AI can support the achievement of the learning objectives or whether certain competences must be specifically learnt and demonstrated without AI. While AI applications can be usefully integrated as a didactic tool, for example in the critical handling of information or in creative processes, there are other skills and competences, especially basic competences, which will also have to be learnt independently (without AI support) in the future. Teachers must therefore reflect on the learning objectives of their course and, if necessary, define them more precisely and consistently align the course and the examination format accordingly.
The aim remains a "constructive alignment" (Biggs & Tang, 2011) between the intended learning objectives, teaching and learning activities geared towards them and suitable examination formats (see also Constructive Alignment | Lehrblick - CUAT University of Regensburg (external link, opens in a new window)).
Implications for examination forms
The key challenge for teachers is therefore to decide whether the use of AI supports or impairs the achievement of learning objectives. If students have to master a skill even without AI, only supervised examination formats can be used, as this is the only way to ensure a legitimate, fair examination environment that leads to meaningful and valid examination results.