Abstract:

This study examines the role of ChatGPT and other AI-powered tools in enhancing teaching and learning in higher education, with particular attention to African contexts. Guided by two objectives, the review analyses how AI tools are pedagogically integrated in higher education and how their use is associated with student engagement and academic achievement. A systematic review and qualitative synthesis were conducted in accordance with PRISMA 2020 and structured using the Population, Intervention, Comparison, and Outcome approach. Peer-reviewed journal articles and conference papers published between 2020 and September 2025 were identified through systematic searches of Scopus, Web of Science, ERIC, and Google Scholar. Studies were included if they reported empirical evidence on ChatGPT or other AI-powered educational tools in higher education settings within Africa or with clear contextual relevance to African higher education, involved students, educators, or institutions, and reported outcomes related to engagement or achievement. Studies were excluded if they were non-peer-reviewed (for example, theses, reports, book chapters), non-empirical commentaries, conducted entirely outside Africa without contextual relevance, not focused on higher education teaching and learning, not available in full text, or not published in English. Following screening, eligibility assessment, and full-text review, 17 empirical studies were retained for synthesis.

The synthesis indicates substantial conceptual and operational variation across the reviewed studies. Student engagement is variously conceptualised as behavioural participation and persistence, cognitive engagement and deep strategy use, emotional engagement and interest, agentic engagement and proactive contribution, social engagement and collaboration, and technology-mediated engagement evidenced through learning analytics such as logins, time-on-task, clickstream activity, and discussion intensity. Academic achievement is operationalised through course grades and GPA, test and examination scores, assignment and rubric-based performance, competence and skills attainment in applied tasks, project outputs and portfolio quality, pass rates and progression, and, in some studies, higher-order outcomes such as problem-solving performance and knowledge-transfer indicators. Overall, AI-powered tools, particularly ChatGPT, are more consistently associated with improvements in engagement-related outcomes than with uniform gains in achievement, reflecting differences in pedagogical design, assessment regimes, and outcome operationalisation. Reported benefits include real-time feedback, adaptive learning pathways, personalised learning support, and scaffolding for complex problem-solving, alongside reductions in instructional workload. Persistent challenges include academic integrity risks, ethical and data privacy concerns, technological dependency, uneven staff capacity, and unequal access shaped by the digital divide.

The study concludes that while ChatGPT and other AI-powered tools hold considerable promise for higher education pedagogy, their educational value depends on pedagogically grounded implementation, institutional readiness, and robust ethical governance. The findings underscore the need for context-sensitive policy frameworks, digital literacy and assessment redesign initiatives, and inclusive access strategies to support responsible and sustainable AI integration in higher education.

Keywords:

ChatGPT, artificial intelligence, higher education, teaching and learning, student engagement, educational technology.