Artificial intelligence in examinations: opportunities and challenges for students!

Artificial intelligence in examinations: opportunities and challenges for students!

On April 29, 2025 took place at the FernUniversity A symposium took place that dealt with the topic of “Testing despite and with AI”. Organized by Dr. Angela Schröder and Andreas Giesbert, both the challenges and the opportunities in dealing with artificial intelligence (AI) in academic teaching were discussed. The two experts advocated a proactive and constructive use of AI tools instead of rejecting them outright.

Schröder emphasized the need to teach students classical scientific methods. This knowledge is essential to be able to correctly select and use important AI tools. She also believes it is crucial that universities give researchers access to AI in order to increase overall research performance. The symposium also highlighted the legal framework and the importance of AI for subject-specific examination services.

Challenges and opportunities

The discussion about the use of AI in teaching is complex. Giesbert pointed out the limits of AI use, especially in introductory modules that ask for basic knowledge. Online exams that are digitally proctored could offer a solution to control the use of aids. Exams should also include simple task forms to ensure equal opportunities.

An example of a form of examination that addresses the use of AI would be the analysis of a text written by ChatGPT. The use of AI tools could be permitted in portfolio exams as long as students document their use. This promotes both creativity and reflection on the use of such technologies. The symposium highlighted the need for educators to be empowered and educated to support the use of AI.

Generative AI in the examination process

As the University of Erfurt As noted, the influence of generative AI on written examination performance depends heavily on the task. While generative AI can create helpful summaries, especially for general topics, there is also a risk of deception and plagiarism.

The design of exam formats should aim to ensure that generative AI is used as a support and not as a substitute. In subjects with empirical projects, AI can help develop survey instruments and evaluate data. Courses can address the use of generative AI and encourage students to reflect on their experiences with the tools.

For example, at the University of Erfurt, an oral examination is planned as a module examination to defend written work. In this case, a grade of 5.0 in one of the two partial examinations is not considered sufficient for the entire examination. This creates another mechanism to ensure the integrity of the exams.

Overall, it is clear that responsible and informed use of AI tools not only changes examination formats, but also opens up new avenues for academic teaching. The discussion about AI in education will continue to be a central topic, also with regard to future developments and challenges in science.