Quiz Generation and Chat Interaction with LLMs: A Design and Implementation of an AI-Based LMS

Large Language Models (LLMs) are increasingly integrated into educational contexts due to their ability to process and generate text through prompt-based instructions. Their adoption enables personalized learning and automated assessment, but also raises concerns about accuracy, bias, and responsible use. In this work, we present CLUE LMS, a Learning Management System that combines standard functionalities, such as access to teaching materials and tests, with LLM-based features. Specifically, the system includes a conversational agent for topic clarification and automatic generation of self-assessment quizzes, both grounded in the teacher-provided materials. The architecture integrates Retrieval-Augmented Generation (RAG) to ensure that generated content remains consistent with course content. To assess its potential, we conducted a user study with four participants. Based on the preliminary results, CLUE LMS is simple to use and effectively supports learning, with the quiz generation feature being the most appreciated. These findings suggest that LLM-based tools, when properly constrained and aligned with educational material, can enhance learning while mitigating common risks associated with generative AI in academic settings.