Information Retrieval System for Automating Quiz Generation and Evaluation Using Large Language Models
Gayatri Kurulkar, Sakshi Ingale, Advait Shinde, Y. N. Jangale, Suhasini A. Itkar · Cureus Journal of Computer Science. · 2025
Artificial intelligence (AI) is playing an increasingly transformative role in education by improving how content is delivered, retrieved, and adapted to learners’ needs. This research presents the design and development of an automated quiz generation and educational question-answering system that combines Retrieval-Augmented Generation with safety mechanisms guided by Guardrails AI, a framework for enforcing content quality and ethical constraints in generative models. The system integrates large language models such as Generative Pre-trained Transformer 4 with LangChain pipelines to retrieve semantically relevant answers from academic documents and generate pedagogically appropriate quiz questions. Tested on 50 university-level queries, the system achieved 92% answer accuracy, 95% semantic relevance, and 96% coherence. These results highlight the potential of combining generative AI with structured retrieval and safety checks to build reliable and responsible educational tools. This work contributes to the growing field of AI in education and opens pathways for future development of scalable, safe, and adaptive learning support systems.