Evaluating the Impact of AI-Supported Adaptive Learning on Spatial Decision-Making in Disaster Planning Education: Evidence from the EPD-Assist Platform

Introduction: Disaster resilience education requires learners and practitioners to interpret geospatial data, assess spatial risks, and make evidence-based planning decisions. However, conventional training approaches generally provide standardized and static content that does not account for differences in users’ prior knowledge, professional backgrounds, or decision-making competencies. This study presents EPD-Assist, an artificial intelligence-supported and GIS-integrated adaptive learning platform designed to strengthen spatial decision-making skills in pre-disaster planning education.

Methods: EPD-Assist combines an AI-supported adaptive learning engine, GIS-based scenario modules, learning analytics, and personalized content recommendations. The platform includes scenario-based tasks addressing earthquake risk planning, flood-risk management, and ecological planning. User interactions, assessment performance, scenario completion patterns, and learning needs are used to adapt content difficulty, scenario sequence, and recommended learning materials. A quasi-experimental pre-test/post-test evaluation framework is proposed to compare the platform with conventional static training. The evaluation will examine learning gain, decision accuracy, scenario completion time, recommendation uptake, user confidence, and perceived usefulness across students, municipal staff, and planning experts.

Results: The proposed framework enables the systematic evaluation of how adaptive learning mechanisms influence spatial reasoning and planning-oriented decision-making. The platform is expected to provide differentiated learning pathways for users with varying expertise levels, while GIS-based scenarios support the application of theoretical knowledge to realistic disaster-planning problems. The learning analytics component also enables instructors to monitor user progress, identify learning gaps, and evaluate decision-making performance.

Conclusions: EPD-Assist offers an integrated framework that connects artificial intelligence, adaptive learning, GIS-based scenario training, and performance monitoring. The proposed approach has the potential to support scalable and personalized disaster resilience education in universities, municipalities, public institutions, and professional training environments.