Emotional engagement in a humor-understanding reading task: an AI study perspective

Emotional engagement is a core facet of overall engagement, influencing our involvement in every task, as our actions inherently evoke emotional responses. With the growing integration of technology in education, it has become more feasible to combine technological devices like eye trackers and artificial intelligence (AI) to gain deeper insights into student engagement in academic tasks. The present study pioneers an exploration into the link between emotional engagement and performance in a humor-understanding reading task through the use of AI to identify emotions linked to task performance. 132 Italian undergraduates took part in a computer-based humor comprehension and appreciation task, which involved completing the Phonological and Mental Jokes task by selecting humorous joke endings and evaluating their funniness. During the task, an AI system assessed participants’ emotional involvement based on facial expressions, distinguishing between neutral state, happiness, and sadness. A positive correlation between measures of happiness and the self-reported perceived funniness was found. Conversely, expressions coded as sadness correlated negatively with the self-reported perceived funniness but positively with the number of correct answers. These results confirm the importance of studying emotional engagement during learning tasks and suggest that expressions of happiness and sadness are differently associated with students’ performance in text comprehension tasks. More broadly, the study provides a model for integrating facial expression detection AI systems to adapt learning tasks to learners’ emotional states.