Educational institutions still rely on manual question paper preparation, which is time-consuming and may lead to inconsistencies in syllabus coverage and difficulty levels. Conventional approaches often face difficulties in maintaining a proper balance among the cognitive levels specified in Bloom’s Taxonomy. Although Generative Artificial Intelligence (GenAI) can automatically generate questions, many existing systems lack proper control over cognitive level, difficulty, and syllabus relevance. This study introduces an AI-driven Question Paper Generation system integrated with Bloom’s Taxonomy alignment. The system takes syllabus topics or keywords as input and generates exam questions using a large language model. A Bloom’s Taxonomy classifier ensures that questions match appropriate cognitive levels, while a difficulty estimation module classifies them into easy, medium, and hard levels. The system also uses semantic similarity analysis to avoid duplicate questions and applies both automated and human evaluation methods to assess quality. The proposed approach aims to generate balanced, syllabus-aligned question papers while reducing educators' manual workload and improving assessment quality.
Waghulde et al. (Tue,) studied this question.