Just ChatGPT it: a mixed methods evaluation of generative AI use among college students

Abstract

The widespread adoption of ChatGPT in higher education has sparked debate, particularly concerning its potential to facilitate academic dishonesty. Despite attempts to tackle ChatGPT use in universities through policy development, obstacles remain because of unclear boundaries concerning its use in more contentious contexts. Our study offers insights into the use of ChatGPT for graded writing assignments from the student perspective, which may help inform future policies regulating ChatGPT use in universities. A simultaneous explanatory mixed-methods approach was used. Study 1 consisted of an online survey (n = 482) of Singaporean university students, while Study 2 consisted of online interviews (n = 20) with Singaporean undergraduates. Guided by the Theory of Planned Behaviour and the Technology Acceptance Model, our survey found that all determinants, except for attitudes, were significant predictors of intention to use ChatGPT for academic purposes. Study 2 also showed students’ mixed attitudes towards the use of ChatGPT. The interviews also found that students defined their ethical boundaries for the academic use of ChatGPT based on several pragmatic considerations and meta-ethical concerns, and their perceptions of ChatGPT's overall impact on their academic work ranged from negative to positive.