A GPT is a GPT, but can it Equalise? A Quantitative Study of Adoption, Inclusion, and Engagement from Students’ Perceptions and Reported Use of Generative AI Tools

Abstract

Despite the growing integration of Generative AI (GAI) tools in higher education, disparities in adoption, engagement, and skill development persist among students. These disparities may be influenced (and potentially further exacerbated) by factors such as socioeconomic status (SES), varying levels of institutional support, and differences in faculty practices (Abbasi et al., 2024; Almassaad et al., 2024; Chung, 2015). This study illustrates the roles of perceived usefulness, socioeconomic status, institutional support, and faculty affiliation in shaping students’ interactions with GAI tools. Data was collected through an anonymised survey completed by 581 consenting students from five faculties at a research-intensive South African university. Through an explanatory quantitative research strategy, statistical tests (including Spearman Rank Correlation, Mann-Whitney U Test, Kruskal-Wallis H Test, and Chi-Square Test), reveal that perceived usefulness significantly drives engagement ( p = 0.531, p < 0.001), while socioeconomic disparities impact engagement frequency (p = 0.0449). Institutional support enhances self-efficacy in reading and writing (p = 0.0303), and engagement positively correlates with skill development outcomes such as writing ( p = 0.2524, p < 0.001). The study was framed by the Adoption, Inclusion and Engagement (AIE) framework developed specifically for this study. Informed by the Technology Acceptance Model, Digital Divide Theory, Self-Efficacy Theory, and underpinned by Socio-Constructivist Learning Theory, the AIE framework highlights the intersection of technology adoption, inclusive practices, and active engagement. Future research should focus on integrating GAI tools into curricula across disciplines, providing guidelines for use, and addressing socioeconomic barriers to access of the tools.

Presenters

Malcolm Roy Weaich
Lecturer, School of Construction Economics and Management, University of Witwatersrand Johannesburg, South Africa

Details

Presentation Type

Paper Presentation in a Themed Session

Theme

2025 Special Focus: Human Learning and Machine Learning—Challenges and Opportunities for Artificial Intelligence in Education.

KEYWORDS

AI, ENGAGEMENT, PERCEIVED USEFULNESS, SOCIOECONOMIC STATUS, SELF-EFFICACY, DIGITAL DIVIDE