Teacher Acceptance and Use of an Explainable AI-Based Student Sustained Attention Monitoring System in Synchronous Online Classrooms

This study presented a student sustained attention monitoring system for synchronous online learning that integrated computer vision with explainable artificial intelligence (XAI) to display real-time behavioral information supporting teachers’ decisions to adjust instruction appropriately. The monitoring system consisted of YOLO-based models to detect facial expressions, blinking rate, eye opening and closure, and signs of fatigue, and employed a Random Forest model to assess sustained attention levels; results were summarized through a user interface designed according to XAI principles that emphasized interpretability and traceability. The sample consisted of 25 teachers from five subject areas who used the system in synchronous online classes for at least a 60-minutes and then completed a user-experience questionnaire. The research instrument was a questionnaire adapted from the System Usability Scale (SUS), covering three aspects: perceived usefulness, ease of use, and user confidence. Descriptive statistics, percentage, mean (x̅) and standard deviation (SD) were used to analyze the data. The findings showed that teachers’ satisfaction with the system was high (x̅ = 4.45, SD=0.20), reflecting the system’s potential to effectively support teacher decision-making in synchronous online instruction and to strengthen confidence in employing digital technologies to enhance teaching and learning.