Artificial Intelligence Adoption and Usage for Academic Writing: A Technology Acceptance Model Perspective in Tanzanian Higher Learning Institutions

Abstract

This study aims to examine the determinants influencing the adoption and continued usage of AI tools for research writing among academicians, focusing on the Tanzania Institute of Accountancy and Mzumbe University (Mbeya Campus). Grounded in the Technology Acceptance Model (TAM), the research investigates the effects of Perceived Usefulness (PU), Perceived Ease of Use (PEU), Behavioral Intention (BI), and Actual Usage (AU) on both adoption and continued usage intentions. A quantitative, cross-sectional design was employed using structured questionnaires distributed to 93 purposively selected academic staff. Exploratory Factor Analysis (EFA) confirmed the reliability and validity of constructs, and multiple regression analysis was used to test the hypothesized relationships. The findings reveal that AU, PEU, and BI are significant predictors of AI tool adoption and continued use, while PU does not have a statistically significant impact. The model explained 62.0% of the variance in adoption and 82.2% in continued usage intention, indicating strong explanatory power. The study concludes that behavioural engagement and tool usability are more critical than perceived usefulness in predicting the sustained use of AI tools in academic writing. It recommends that institutions enhance AI training, promote ethical usage, integrate AI into academic systems, and address user experience concerns to foster widespread and responsible AI adoption in Tanzanian higher education.