Researchers at the HJ Institute of Technology and Management (HJITM) have published a study analyzing how university students’ “guilt” over using generative AI affects their usage behavior and word of mouth (WOM) in Ethics & Behavior (SSCI), an international journal in the field of ethics and behavior.

The researchers surveyed 296 university students in Korea and analyzed the data using partial least squares structural equation modeling (PLS-SEM). They examined how the guilt students feel when using AI for assignments intervenes in the process by which ChatGPT’s characteristics — perceived intelligence, knowledge acquisition, novelty and customization — lead through utilitarian and hedonic benefits to use and recommendation.

The analysis showed that perceived intelligence of the AI (β=0.290) and knowledge acquisition (β=0.421) increased utilitarian benefits, while novelty (β=0.400) and customization (β=0.341) increased hedonic benefits, the enjoyment of using it. Both types of benefit led to use and recommendation, with hedonic benefits having the strongest effect on both use (β=0.408) and recommendation (β=0.408).

The role of guilt stands out. Guilt lowered both utilitarian benefits (β=−0.134) and hedonic benefits (β=−0.239), and directly reduced willingness to recommend the tool to others (β=−0.207). However, it had no significant effect on actual use. One possible interpretation is the emergence of “quiet use,” in which students keep using AI despite feeling uneasy but do not reveal or recommend it.

The study is significant in combining the moral emotion of guilt with a benefit-centered model to explain the acceptance of generative AI. It suggests that rather than banning AI outright or leaving it unaddressed, universities should provide clear guidance on what is permitted and on proper citation and use, creating an environment in which students can use AI openly and share their experiences.

HJITM continues to conduct research on the use of generative AI in education and related ethical issues.