Teaching Artificial Intelligence (AI) at scale presents unique challenges in designing effective, learner-centered educational experiences. This paper explores how data-driven feedback from learners in a wide-scale platform is used as a reference in mapping the strengths and weaknesses perceived by learners in Massive Open Online Courses (MOOCs) on AI. Using responses based on post-course surveys, we use statistics and Latent Dirichlet Allocation (LDA) to uncover recurring themes in learners’ choices and open-ended comments. The findings reveal that different demographics have different course perceptions, and in general, integrating interactive content is one of the attractions to increase learners’ engagement in completing their courses. These insights offer the opportunities to refine MOOC structure and content for a more inclusive and engaging learning environment. This study contributes to the growing body of research in AI education, especially in large-scale online formats, by demonstrating how analysis of user feedback can pinpoint actionable insights. The results of this exploratory study can provide recommendations to course designers for preparing AI education content across different user segment profiles to continuously improve AI-focused MOOCs and better meet the evolving needs of learners.