Researchers at the HJ Institute of Technology and Management (HJITM) have published a study analyzing the factors that determine university students' actual use of ChatGPT in Information Development (SSCI), an international journal in information science.

Drawing on the theory of planned behavior (TPB), the researchers analyzed responses from 232 Korean university students using structural equation modeling. They examined whether perceived intelligence of ChatGPT, knowledge application, and usability lead through attitude to intention to use and actual use, and what roles are played by subjective norms (the expectations of people around the user), perceived behavioral control, and trust.

The analysis showed that perceiving ChatGPT as intelligent sharply increased knowledge application (β=0.676) and also raised usability (β=0.400), attitude (β=0.387), and trust (β=0.185). Attitude was shaped by perceived intelligence along with knowledge application (β=0.229) and usability (β=0.197).

The factor with the greatest influence on intention to use was subjective norms, the sense that people nearby expect one to use it (β=0.504). Attitude (β=0.256) and trust (β=0.190) also increased intention to use, but perceived behavioral control, the sense that one can easily use it oneself, had no significant effect. Intention to use translated strongly into actual use (β=0.601).

The study is notable for explaining university students' use of generative AI by combining the theory of planned behavior with AI-specific characteristics: perceived intelligence, knowledge application, and trust. It suggests that in education it is important to share how instructors and peers use the tool, and to teach students how to judge the reliability of information produced by AI.

HJITM continues its research on the use of generative AI in education.