A Comparative Analysis of Artificial Intelligence Learning and In-Class Learning in Kenyan Universities

Abstract

This study presents a comparative analysis of Artificial Intelligence (AI) learning and traditional in-class learning in Kenyan universities, focusing on learning outcomes, student engagement, and accessibility. The study was conducted across six universities—three public and three private—and the study involved 424 undergraduate students and 35 instructors, utilizing a mixed-methods approach. Quantitative data from surveys and performance records were complemented by qualitative insights from interviews and focus group discussions. The results revealed that students using AI learning platforms outperformed those in traditional settings, particularly in Science, Technology, Engineering, and Mathematics (STEM) disciplines, with AI learners achieving an average score of 78.4% compared to 72.9% for in-class learners. Engagement levels were also higher among AI learners, who reported more frequent interaction with learning materials and instructors, especially during asynchronous sessions. However, significant challenges were identified in terms of accessibility, especially in rural universities where unstable internet and a lack of digital infrastructure hindered AI adoption. In contrast, in-class learning was more accessible in these regions but faced issues of overcrowded classrooms and limited resources. The study suggests that while AI learning offers significant advantages in terms of personalization and flexibility, a blended learning model combining both AI and in-class methods may provide the most effective educational experience. This model could address the strengths and limitations of each approach, especially in addressing the digital divide and fostering critical thinking. The findings suggest policy reforms to improve digital infrastructure, promote instructor training, and ensure inclusive access to AI-driven education in Kenyan universities.