This page contains automatically translated content.

08/20/2025 | Press Release

AI in biology lessons: study sees only limited learning success and warns of risks

The use of ChatGPT in biology lessons only minimally increases the learning success of pupils on the subject of evolution, but at the same time leads to a more positive attitude towards AI-supported learning tools. This is shown by a new study in which Dr. Tim Hartelt, research associate at the Department of Didactics of Biology at the University of Kassel, was involved. At the same time, the researchers warn of the risks of using such technologies unthinkingly and unilaterally.

Dr. Tim Hartelt from the Department of Didactics of Biology at the University of Kassel.Image: Private.
Dr. Tim Hartelt from the Department of Didactics of Biology at the University of Kassel.

Two groups were compared in the study with 241 upper secondary school students: An experimental group used ChatGPT to deepen their knowledge of evolution, while the control group used traditional internet searches. The study design included before, after and follow-up tests over a period of around two months.

The results show that the ChatGPT group achieved only a slight improvement in their evolution expertise . No significant differences were found between the groups in terms of digital skills, particularly in the critical use of AI. What was striking, however, was the change in attitude: students who worked with ChatGPT expressed a greater willingness to use AI tools regularly in the future - regardless of the actual increase in performance.

The researchers see this as a potential risk. A positive attitude combined with a low learning effect could lead to increased use of ChatGPT without significantly increasing learning success. "It is crucial that students develop skills in dealing with AI, for example in the targeted formulation of input (prompt engineering), in order to be able to use AI tools effectively," says Hartelt. He therefore advocates a reflective, critical use of such tools in the classroom and the teaching of appropriate evaluation skills.

The results of the study by Helena Aptyka (University of Cologne), Jörg Großschedl (University of Cologne) and Tim Hartelt (University of Kassel) were published in the International Journal of Science Education and are available online at: https://doi.org/10.1080/09500693.2025.2524083.