Digital learning innovation: engineering students’ learning motivation for AI scaffolding

Abstract This study compares engineering college students’ learning motivation between AI-featured and traditional scaffolding methods using a mixed-methods approach. Two different types of class activities, one involving solving a computational problem and the other assembling an essay, were designed to be carried out using both AI-featured and conventional methods. All activities were designed using the Attention, Relevance, Confidence, and Satisfaction (ARCS) model and results were analyzed using both quantitative and qualitative means. Findings show that participants’ motivation was significantly higher in all Attention, Relevance, Confidence, and Satisfaction categories when they used AI-featured methods for solving computational problems while their motivation was significantly higher when they used AI-featured methods only in Attention and Confidence for the essay activity. Qualitative findings describe how AI features motivated learners based on the different types of scaffoldings. Additionally, how the AI features are related to each ARCS model categories are summarized. The results suggest strategies for optimizing the design of AI-featured scaffolds to motivate learners in STEM fields grounded in the ARCS model. This involves linking learners’ goals and interests to real scientific applications of AI by expanding ‘Relevance’ and adjusting learning requirements to strengthen ‘Confidence’ and ‘Satisfaction’.