PREDICTING MOBILE LEARNING APPS’ USER RATING BASED ON THEIR USABILITY EVALUATION BY GENERATIVE ARTIFICIAL INTELLIGENCE (AI)

This article explores the prediction of mobile learning apps’ user ratings from various online app stores by means of the apps’ usability evaluation by generative artificial intelligence (AI). The generative AI robot employed in the study was Meta Llama, which was requested to award rating scores to the eight major usability dimensions, namely, (1) content/course quality, (2) pedagogical design, (3) learner support, (4) technology infrastructure, (5) social interaction, (6) learner engagement, (7) instructor support, and (8) cost-effectiveness of 17 currently popular mobile learning apps. The apps’ user ratings were basically the average of those “star” ratings from the two leading online apps stores, viz., Apple App Store and Google Play. Multiple regression of such a user rating (as the dependent variable) on the Llama rating scores for the eight dimensions or their subset (as the independent variables) was performed. It was found that multiple regression of such a user rating on the Llama rating scores for the five usability dimensions (3) learner support, (4) technology infrastructure, (5) social interaction, (7) instructor support, and (8) cost-effectiveness constituted the final prediction model (R² = .596, the F-test’s F statistic = 2.947 with its p-value = 0.069 < 0.1) with the respective regression coefficients for the above five Llama rating scores being -2.150, .355, 2.578, -2.502, and .223 and the corresponding t-tests’ p-values being .008 < 0.05, .729, .043 < 0.05, .238, and .669.