Deriving Design Principles for AI-Adaptive Learning Systems: Findings from Interviews with Experts

Abstract

AI applications are increasing in the field of education, from laboratory set-ups to contemporary and complex learning systems. A great example of such systems is AI-enabled adaptive learning systems (AI-ALS) that promote adaptive learning. Despite its promised potential, there are challenges such as design issues, highly complex models, and lack of evidence-based guidelines and design principles that hinder the large-scale adoption and implementation of AI-ALS. The goal of this paper thus is to establish a set of empirically grounded design principles (DPs) of AI-ALS, that would serve well in a university context. 22 interviews were con-ducted with experts knowledgeable about the design and development of AI-ALS. Several rounds of coding and deep analysis of the expert interviews revealed features and functionalities of AI-ALS; purposes for designing and using AI-ALS; and recommended improvements for AI-ALS as requirements. These requirements were translated to 13 preliminary DPs. The findings of this study serve as a guide on how to better design AI-ALS, that will improve the learning experiences of students.