Swiss universities move beyond traditional exams as AI reshapes study

Artificial intelligence is changing university study, prompting the University of Zurich and Zurich University of Applied Sciences to rethink how students’ skills are assessed, SRF reported. The institutions say AI can make work easier but cannot replace understanding and expertise.

Students can now create texts, organise ideas and generate summaries in seconds. But universities warn that relying on AI to do the thinking risks preventing students from developing the required competence.

“The danger is that students will let the technology do the thinking for them,” said Gabriele Siegert, vice-rector of the University of Zurich.

Both institutions are placing less emphasis on the traditional bachelor’s or doctoral thesis as the sole proof of achievement. Instead, they want to assess the process behind the final product. A well-written seminar paper does not necessarily show that a student has understood the subject.

The universities are therefore using more oral examinations, presentations and defences of bachelor’s theses. Students should be able to explain their arguments, justify decisions and critically question their results.

Lisa Messenzehl, head of the Teaching Technologies and Didactics group at the Zurich University of Applied Sciences, said subject knowledge must still be available without technological assistance. A physiotherapist, for example, must know human anatomy without using a chatbot.

Siegert expects oral examinations and presentations to become more common. However, such formats are difficult to organise for courses with between 800 and 1,000 students. Written multiple-choice examinations may therefore continue to be held on site in a protected environment where AI cannot be used.

The Zurich University of Applied Sciences has decided against using AI detectors to prevent cheating. Messenzehl said the systems were not reliable enough.

Both universities nevertheless consider AI skills an important competence for the future. People who cannot use the tools may face disadvantages in working life, but AI does not replace the knowledge needed to become a physiotherapist, engineer or lawyer.

AI can make work easier, the institutions say, but students must still learn, understand and develop their skills themselves. A convenient AI-based diploma remains an illusion.