RESEARCH ON BUILDING AI CHATBOT TO SUPPORT TEACHING

Abstract

This article presents a solution for building a chatbot using Artificial Intelligence to support the teaching process, with the goal of optimizing teachers' workloads and improving the quality of learning. The chatbot was developed by combining advanced technologies such as Natural Language Processing, Large Language Models, and Retrieval-Augmented Generation. These technologies enable the chatbot to understand context, handle complex professional questions, and provide quick, accurate responses.

Experimental results show that the chatbot performs with high efficiency in providing correct and natural answers, reflecting its ability to handle various types of questions. The system has proven to be a reliable tool, assisting teachers with tasks such as lesson planning, information retrieval, and answering student inquiries. The chatbot's responses were evaluated as coherent, easy to understand, and appropriate for an educational context.

However, the article also points out some limitations that need improvement, including an incomplete ability to understand complex contexts and a limited scope of knowledge. Future research will focus on upgrading the Natural Language Processing model, expanding the knowledge base, and integrating multimedia features such as voice and image recognition. The ultimate goal is to transform the chatbot into a smarter and more comprehensive educational support tool, contributing to the promotion of personalized learning experiences in the digital age.