The Effect of AI-Generated Emotional Pictures and the Role of Working Memory on Vocabulary Learning

Drawing on the Cognitive-Affective Theory of Learning with Media (CATLM), this study explores how emotional design in multimedia instruction impacts vocabulary acquisition among English as a foreign language (EFL) learners. Considering working memory (WM) constraints, the research assesses the effects of Artificial Intelligence (AI)-generated pictorial stimuli with varying emotional loads—negative, neutral, and positive—on learning outcomes. Thirty-six vocabulary items, paired with emotionally differentiated pictures, were presented to elementary-level learners (n = 348), with learning gains measured via pretest-posttest comparisons. Working memory capacity (WMC) was evaluated using the Reading Span Task. Results indicated that emotional load significantly influenced vocabulary retention, with positive and negative pictures yielding the strongest effects. While WMC predicted learning for positively and neutrally loaded words, its impact diminished for negatively loaded items, supporting CATLM’s assertion that affective factors can supersede cognitive capacity in learning. These findings highlight the interplay between emotional engagement and cognitive processing, emphasizing the need for instructional materials that balance both elements. The study contributes to the understanding of the role of affective design and AI-generated visuals in optimizing language learning, offering practical implications for multimedia-based EFL pedagogy.