
Syddansk Universitetsbibliotek - LibGuides


In a field study by Bastani et al. (2024), approximately 1000 students in a Turkish high school were divided into three groups in the subject of mathematics:
During the exercises, the first group improved by 48%, and group 2 improved by 127%. When the tools were removed for the exam, group 1 performed 17% worse than group 3 (the control group), who had not used AI at all. And group 2 performed approximately as well as the control group.

(Graphics generated by Copilot based on the article from Bastani et al., 2024)
So according to this study, if you use AI uncritically, you risk understanding the material less than if you had not used AI. If you use AI equipped with some kind of tutor or pedagogical limitation, you perform just as well as if you had not used AI - but you solve the tasks faster. And any improvement during the semester will not necessarily be transferred to the exam.
So get used to using AI as a sparring partner for what you have to learn. Don't leave all your thinking capacity to AI, and don't use it as a checklist.


University Library of Southern Denmark Odense | Esbjerg | Kolding | Slagelse | Sønderborg +45 6550 2100 | sdub@bib.sdu.dk


