AI Use and Academic Persistence among Undergraduate Students in Tanzania

Abstract The use of Artificial Intelligence (AI) tools such as ChatGPT and Google Gemini has become a routine part of student life in higher education worldwide, yet very little is known about how students in Sub-Saharan Africa actually experience these tools, or how that experience relates to whether they stay in or leave their studies. This qualitative study examines how undergraduate students at the Tanzania Institute of Accountancy (TIA) use AI in their learning, how AI helps them understand difficult material, how AI-supported learning relates to their intentions to persist in or withdraw from their studies, and what challenges they encounter along the way. Data were generated from 104 undergraduate students in the Bachelor of Public Sector Accounting and Finance (BPSAF) and Bachelor of Human Resources Management (BHRM) programmes through a semi-structured, Swahili-language interview guide, and analysed using Braun and Clarke's six-phase thematic analysis framework. Students described active, often critical engagement with AI tools across three distinct patterns of use, and identified four specific ways AI helped them understand difficult content, including two mechanisms - bilingual Swahili-English explanation and unlimited, judgement-free follow-up - that extend the existing literature on AI in higher education. Most students (86%) said AI made learning easier, and most (68%) had never considered withdrawing from their studies; among those who had, AI was frequently described as a reason they stayed. At the same time, a smaller group of students described AI use as undermining rather than supporting their learning, including cases of examination failure linked to uncritical reliance on AI-generated answers. Five categories of challenge were identified: AI inaccuracy, unreliable or unaffordable internet access, subscription paywalls, cognitive over-reliance, and a newly identified category of direct academic harm. This study offers one of the first qualitative accounts of AI use and academic persistence from a Tanzanian higher education setting, and contributes context-specific evidence to a literature that has so far drawn almost entirely on research from outside Africa.