Uswatun Hasanah, Haikal Abror, Syahroni Hidayat, Saiful Ridlo, Taofan Ali Achmadi, Ahmad Sehabuddin
The rapid advancement of digital technology has outpaced the preparedness of the education sector, resulting in significant gaps that may hinder effective teaching and learning. Without timely and optimal integration of technology, particularly in instructional and assessment processes, teachers may struggle to meet evolving learning demands, and students risk losing access to more relevant and efficient learning methods. This study aims to develop a web-based automatic essay question generator powered by the Large Language Model Gemini 1.5 Flash. The system is designed to assist educators in generating cognitively aligned essay questions based on Bloom's taxonomy (C1-C6), tailored to the content of instructional modules. A quantitative experimental research design was employed to evaluate the system. Validation by subject-matter experts yielded an Aiken's V score of 0.85, indicating a very high level of content validity. Additionally, teacher feedback demonstrated strong agreement regarding the system's quality and utility, with an acceptance rate of 80.3 %. This research serves as an initial step toward the practical integration of large language model (LLM) technologies in education, offering recommendations for their application in automated question generation systems for instructional purposes. © 2025 IEEE.
Universitas Negeri Semarang, Department of Computer Engineering, Indonesia; Universitas Negeri Semarang, Department of Informatics and Computer Engineering Education, Indonesia; Universitas Negeri Semarang, Department of Electrical Engineering, Indonesia; Universitas Negeri Semarang, Department of Natural Science Education, Indonesia; Universitas Negeri Semarang, Department of Family Welfare Education, Indonesia; Universitas Negeri Semarang, Department of Economics Education, Indonesia
