Perceptions of AI-based learning tools among Tanzanian university students; systematic PRISMA review

Abstract

Large language model chatbots such as ChatGPT are reshaping higher education pedagogy in Tanzania, where institutions face infrastructural constraints, limited AI governance, and digital inequalities that shape student engagement with AI tools. This systematic review synthesises empirical studies on Tanzanian university students’ perceptions, adoption patterns, and concerns about academic integrity in relation to AI-based learning tools within the broader Sub-Saharan African higher education context. Following PRISMA 2020 guidelines, searches across six databases identified 525 records, with 38 empirical studies published between 2018 and 2025 included after screening and quality appraisal using MMAT. Findings show that awareness of AI tools among Tanzanian university students ranges from 77% to 93%, with ChatGPT being the most widely used tool and an acceptance rate reaching 85.7%. Students mainly use AI for academic writing (47.5%) and idea generation (38.2%), while perceived ease of use strongly predicts adoption (marginal effect 0.1007). Reported benefits include personalized learning, improved productivity, and enhanced access to learning content. Key challenges involve weak ICT infrastructure, high mobile data costs, limited AI literacy, and absence of clear institutional AI policies. Concerns about academic integrity remain significant, as higher AI use and perceived usefulness are associated with increased integrity violations. Overall, Tanzanian students show positive attitudes towards AI tools, but usage is constrained by connectivity barriers, low AI literacy, and weak governance. Policy recommendations emphasize infrastructure investment, AI literacy development, institutional regulation, and future longitudinal research including rural and underserved populations to ensure equitable and responsible AI integration in higher education.