Abstract:

The rapid expansion of generative artificial intelligence has introduced significant changes in higher education, particularly through the increasing use of tools such as ChatGPT in students’ academic work. These tools provide new opportunities for learning support, faster access to information and assistance in academic tasks. At the same time, their growing presence raises important questions regarding students’ competencies, patterns of use and perceptions of ethical acceptability when artificial intelligence is integrated into academic activities.

This study examines the role of ChatGPT in higher education with a focus on students’ self-assessed competence in using the tool, the frequency and manner of its use, and students’ perceptions of its usefulness and ethical acceptability in academic work. The research was conducted among first-year undergraduate students enrolled in the Information and Communication Technologies course at the Zagreb School of Economics and Management during the 2024/2025 academic year.

A quantitative research design was applied using pre- and post-survey instruments administered to the same group of students. The final sample consisted of 104 respondents who completed both surveys and participated in a course assignment that required writing an essay with the support of ChatGPT. The assignment served as a structured learning activity allowing students to gain practical experience with the tool in an academic context. Data were analyzed using descriptive statistical methods and correlation analysis in order to identify changes in students’ perceived competence and relationships between usage patterns and attitudes toward the tool.

The results indicate a statistically significant increase in students' self-assessed competence in using ChatGPT after completing the assignment. The average competence score increased from M = 3.52 before the assignment to M = 3.94 after the assignment, with the difference being statistically significant (t(103) = 4.44, p < 0.001, Cohen's d = 0.44), indicating a moderate effect size. The structured integration of the tool within the course therefore contributed to higher levels of confidence and perceived ability among students. However, the frequency of ChatGPT use was not significantly related to students' perceived competence Instead, the manner in which students used the tool was more strongly related to their perceptions of usefulness and ethical acceptability. Students who used ChatGPT primarily for understanding concepts, generating ideas and structuring academic work reported higher levels of perceived usefulness and ethical acceptability compared to students who relied on the tool primarily for generating ready-made content.

The findings highlight that the educational impact of generative artificial intelligence depends less on how often such tools are used and more on how they are integrated into the learning process. Structured and guided use of artificial intelligence tools can support the development of digital competencies while encouraging responsible and reflective use of technology in higher education.

Keywords:

Generative artificial intelligence, ChatGPT, higher education, student competence, ethical use.