Related Experiment Video
Updated: Jun 23, 2026
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
Exploring Medical Students' Perceptions Regarding ChatGPT and AI Studying at the University of Niš: A Study on Usage,
Aleksandra Ignjatović1,2, Marija Anđelković Apostolović1,2, Lazar Stevanović3
1Faculty of Medicine, Department of Medical Statistics and Informatics, University of Niš, Niš, Serbia.
Objectives:
The primary objective of the study was to analyze the extent of ChatGPT usage and students' attitudes about ChatGPT. The secondary objective is to examine the potential underlying structure of students' perceptions about this AI tool in academic settings. Additionally, the focus on students studying in Serbian and English may shed light on the cultural and linguistic influence of technology implementation in educational settings.
Methods:
This descriptive study investigated how medical students perceive the application of ChatGPT in educational settings using a 3-part questionnaire: a demographic section collecting participants' overall experience and perceptions of ChatGPT, and a section collecting students' attitudes toward artificial intelligence who did not use ChatGPT.
Results:
In the survey, there were 1212 students (1037 students in the Serbian and 175 students in the English language). Almost four-fifths of students have heard of ChatGPT (79.4%), but less than half of students have used ChatGPT (42.2%). Principal component analysis revealed 4 components: Positive Perception and Usefulness, Negative Impacts and Risks, ChatGPT Usability and Improvement, and Interaction and Communication Challenges items.
Discussion:
A positive attitude was prevalent in our population; however, students were also aware of the limitations of ChatGPT, which were recognized as potential academic dishonesty and a risk of job loss. There were significant differences in students' agreement with various statements across 3 components among students studying Serbian and English, implying a possible role of linguistic barriers in AI responses. Therefore, using ChatGPT to promote education and healthcare should be done ethically and responsibly, considering the possible improvements in AI tools and the risks and issues they raise.
