Development of an AI-Driver Digital Platform to Support STEM Instructional Design and Learning for Teachers

Abstract

This research is a research and development (R&D) study aiming to 1) examine the current conditions, problems, and needs of teachers in implementing STEM-based learning; 2) design and develop an artificial intelligence-driven digital platform (AI-driven platform) to support the design of STEM learning management; and 3) investigate the effectiveness of the platform and its implementation guidelines. The research was conducted in three phases. The first phase involved studying the context and needs of 30 teachers using questionnaires and in-depth interviews. The second phase focused on designing the conceptual framework and developing the EDSTEM platform using prompt engineering and cloud infrastructure technology, followed by expert evaluations. The third phase evaluated the platform's effectiveness through assessments by three experts in relevant fields and feedback from 30 actual users. The research findings revealed 1) The current state of STEM learning management was at a moderate level, 3.31. Teachers showed high confidence in instructional management (3.97) and activity design (3.77), while weaknesses were found in assessment practices, understanding of STEM concepts, and institutional support. 2) The platform’s conceptual framework was evaluated as highly consistent with theoretical foundations 4.53, and three main tools were developed: a lesson plan generator, an assessment tool, and an AI assistant. 3) The EDSTEM platform demonstrated high efficiency in terms of content, design, and functionality, 4.40–4.58, effectively addressing teachers' needs in STEM education.

Keywords: AI-driven, digital platform, STEM-based learning, learning management design