Incident factors in the use of ChatGPT and dishonest practices as a system of academic plagiarism: the creation of a PLS-SEM model

Abstract

The objective of this study was to construct a causal model that explores the factors influencing university students’ behavioral intention to use ChatGPT as a learning resource, and to understand the reasons behind their engagement in dishonest practices when incorporating ChatGPT into their academic work. We gathered data through a survey, with the participation of 368 university students. Our analysis employed a causal model based on the Partial Least Squares Structural Equation Modeling (PLS-SEM) approach. On one hand, we verified the hypotheses regarding factors contributing to the intention to use ChatGPT, such as the quality of information generated by the software, the credibility of this information, and student satisfaction with the responses provided. On the other hand, we found compelling evidence of significant factors affecting the intention to use ChatGPT inappropriately, including the absence of clear regulations on plagiarism and corresponding penalties in universities, the students’ insufficient research and academic skills (such as conducting research or writing in a scholarly manner), the adverse impact of teachers’ workload (excessive tasks or insufficient commitment to assessments), and the general lack of interest and motivation towards academic tasks. Collectively, these factors accounted for 53.20% of the variance in students’ behavioral intention to utilize ChatGPT. These findings affirm the effectiveness of this model in explaining the software’s role in text production and quality, as well as students’ tendencies toward dishonest use.