Original scientific paper
https://doi.org/10.15516/cje.v27i3.6596
Appropriate and Inappropriate Use of Artificial Intelligence Language Models by Students: Teachers’ (Self-)Perceptions and Experiences Across Educational Levels
Irena Miljković Krecar
orcid.org/0000-0003-2167-4947
; VERN’ University
Nina Pavlin-Bernardić
orcid.org/0000-0002-8194-5668
; University of Zagreb Faculty of Humanities and Social Sciences
Abstract
This study examined primary, secondary, and higher education teachers’ self-efficacy in identifying AI-generated student work, concerns about potential negative effects of AI on learning, perceptions of the frequency of appropriate and inappropriate AI use among students, and actual AI-giarism detection. A total of 511 Croatian teachers completed an online survey comprising newly developed scales. Welch’s ANOVA revealed differences between educational levels for most variables, with small to small-to-medium effect sizes. Secondary school teachers perceived the highest rates of inappropriate AI use and had higher confidence in detecting students' AI use than higher education teachers, while primary school teachers reported the lowest rates of appropriate use and detection frequency. Academic AI concern was high across all groups, and self-efficacy scores were slightly above the midpoint of the scale. Hierarchical regression analyses showed that the perceived inappropriate AI use was predicted by the extent of teachers’ AI tool use and concerns, whereas the actual AI-giarism detection frequency was predicted by education level, recognized AI tools, extent of AI use, and confidence in detection. The findings highlight the need for targeted teacher training, clear institutional guidelines, and AI-resilient assessment design.
Keywords
academic AI concern; appropriate AI use; inappropriate AI use; self-efficacy; teachers’ perceptions; AI-giarism detection
Hrčak ID:
345116
URI
Publication date:
29.9.2025.
Visits: 323 *