AI and Academia: Navigating the Adoption of Artificial Intelligence in Universities

Abstract

This study investigated the utilization of artificial intelligence (AI) platforms in Bangladeshi higher education institutions, with an emphasis on determining which AI platforms are most widely used and assessing the variables that affect AI adoption. The particular criteria influencing platform preferences and the comparative analysis of various types of institutions remain unclear despite the abundance of research on AI adoption. The study employed logistic regression analysis to test two primary hypotheses: (1) private universities had a higher level of AI adoption compared to public universities, and (2) better technological infrastructure was positively associated with higher AI adoption levels. Data was collected through a structured survey administered to a representative sample of 100 participants from various higher education institutions, capturing information on AI adoption levels, institutional type, technological infrastructure, technical expertise, and financial constraints. The analysis revealed that the most widely adopted AI platforms were ChatGPT, followed by AWS, Google Cloud Platform, and Google Gemini. The logistic regression results supported the hypotheses, indicating that private universities were more likely to adopt AI at higher levels compared to public institutions. Additionally, better technological infrastructure was associated with higher AI adoption levels. The confusion matrix demonstrated that while the model performed well in predicting AI adoption levels, there were some misclassifications, particularly between high and medium adoption categories.