An In-Depth Examination of the Challenges Facing GAI Engagement and Student Usage Trends and Their Impact on Teaching and Learning

First presented locally at the University of the Witwatersrand, Johannesburg, for the GAI Project Workshop in March 2025, subsequently to be presented at the Thirty-Second International Conference on Learning July 08 - 10, 2025 at the University of Granada, School of Education, Granada, Spain. The presentation provides a comprehensive analysis of the multifaceted challenges influencing the engagement of GAI Tools within higher education, specifically at a research-intensive African university. The study employs robust quantitative methodologies, including the Spearman's Rank Correlation, Mann-Whitney U tests, Kruskal-Wallis H tests, and the Principal Component Analysis (PCA), to elucidate the critical factors shaping interaction with GAI technologies. Key findings reveal a statistically significant positive correlation (p = 0.524; p < 0.0001) between familiarity with GAI tools and student engagement levels. This highlights that institutional strategies aimed at increasing GAI adoption must prioritise enhancing familiarity, perceived usefulness, and accessibility through structured training and awareness initiatives. The research further identifies socioeconomic status (SES) as a pivotal determinant, where self-funded students exhibit higher engagement with AI tools compared to financially aided peers (p < 0.05). This underscores the persistence of the digital divide, necessitating targeted interventions to support equitable AI integration across socio-economic demographics. The analysis also demonstrates that faculty affiliation significantly influences both perceived usefulness (PUS) and engagement (ENG) with GAI tools, as evidenced by Kruskal-Wallis H test results (PUS: H = 11.58, p = 0.0207; ENG: H = 38.44, p < 0.05). These outcomes stress the importance of faculty-specific approaches to AI readiness, ensuring balanced adoption across disciplines. PCA results indicate that institutional support impacts students’ self-efficacy holistically, with Principal Component 1 explaining 60.1% of variance. This depicts that enhancing AI-related institutional support can broadly improve academic confidence across critical skills such as reading, writing, critical thinking, and problem-solving. This body of work contributes to ongoing discourse on digital equity, technology acceptance, and pedagogical transformation in the era of AI-driven education. It provides actionable insights for policymakers, educators, and institutional leaders aiming to foster inclusive and effective AI integration within teaching and learning environments. This research is currently under consideration for publication, where it will contribute to scholarly debates on the intersection of AI, education, and socio-economic dynamics in global higher education contexts. The presentation will be available for downloaded once the article linked to this presentation has been published.AcknowledgementsFunding for this research was provided by the Department of Higher Education and Training (DHET) through the University Capacity Development Programme (UCDP) which funded the Generative Artificial Intelligence (GAI) project at the University of the Witwatersrand (Wits). © GAI Project at Wits University.