The Last Decade of Automatic Question Generation: A Review of Techniques, Limitations, and Applications in Business Process Management Education

Automatic Question Generation (AQG) is a research area that employs Natural Language Processing (NLP) techniques to automatically generate questions from textual content. Although it is still considered an emerging field, AQG has experienced significant growth in recent years, driven by advances in artificial intelligence, especially in deep learning and large language models, as well as by the increasing demand for scalable educational technologies. This article presents a Systematic Literature Review (SLR) focused on AQG research conducted over the last decade. The review aimed to identify and analyze the main computational approaches, practical applications, existing limitations, evaluation methods, and the degree of acceptance by education professionals. The SLR was carried out using major academic databases, resulting in the selection of 103 relevant studies, of which 90 are original research articles and 13 are literature reviews. The results show a clear trend toward the adoption of Transformer-based models, which have significantly improved question generation quality. However, the analysis also reveals a lack of consensus regarding standardized evaluation metrics, particularly for automatic assessments, and a notable gap in studies that investigate how educational professionals perceive and accept questions generated by these systems. This highlights an important area for future research.