UNIVERSITY STUDENTS’ PERCEPTIONS OF CHATGPT: FACTORS THAT AFFECT THEIR BEHAVIOURAL INTENTION, USAGE AND UTILITY PERCEIVED
O. Silva1, J. Nunes2, Á. Sousa3
ChatGPT is a generative artificial intelligence (AI) tool with the potential to transform education by enhancing student engagement and making learning more personalized and interactive. However, concerns exist within the educational community regarding its potential for facilitating plagiarism and undermining independent thinking. Therefore, it is crucial to thoroughly investigate both the benefits and drawbacks of using ChatGPT, particularly among higher education students, to implement appropriate measures.
Given young people's strong appetite for technology, it is in the best interest of higher education institutions and their stakeholders to establish best practice guidelines for using ChatGPT. Evaluating the key factors influencing behavioural intentions towards ChatGPT based on higher education students' perceptions is essential.
This study examines the factors [Performance Expectancy (PE); Effort Expectancy (EE); Social Influence (SI); Facilitating Conditions (FC)] that influence university students' behavioural intentions (BI) to use ChatGPT 4.0 (free version) and its actual use (GPTU), based on a model adapted from the Unified Theory of Acceptance and Use of Technology (UTAUT1 and UTAUT2).
The primary objective of this study is to determine whether technology use factors significantly affect behavioural intentions towards ChatGPT and its usage. Additionally, the study tests whether using ChatGPT impacts perceptions of ease of use (PEU) and perceived usefulness (PU) and whether perceived trust (PT) moderates the relationship between the constructs (PE, SI, EE, FC) and behavioural intention (BI).
The study was conducted at a Portuguese higher education institution with 585 students, using a questionnaire containing 32 items divided into 9 constructs [PE, EE, SI, FC, BI, GPTU, PU, PEU, and PT] and sociodemographic variables.
Data were analysed using Partial Least Squares Structural Equation Modelling (PLS-SEM), fulfilling all reliability and validity (convergent and discriminant) requirements for the constructs. Most hypotheses were confirmed by significant results, with social influence (SI) and facilitating conditions (FC) being the most influential factors for behavioural intention (BI) to use ChatGPT in learning. The use of ChatGPT (GPTU) significantly impacts perceptions of ease of use (PEU) and perceived usefulness (PU). Perceived trust moderates the relationship between the constructs (PE, SI, FC) and behavioural intention (BI).
The findings can assist policymakers in higher education institutions in developing effective and efficient policies to enhance the use of ChatGPT in Portuguese higher education.
Keywords: University students’ perceptions, ChatGPT, data analysis, PLS-SEM.