Educational chatbots have gained prominence as support tools for teaching programming, particularly in introductory learning contexts. This paper presents a Systematic Mapping Study (SMS) that investigated how such agents have been developed and applied in programming education. From an initial set of 3,216 publications, 54 studies were selected and analyzed based on five research subquestions, addressing chatbot types, programming languages used, educational content covered, interaction models, and application contexts. The results reveal a predominance of chatbots designed for Python instruction, focusing on fundamental programming concepts, and employing a wide variety of pedagogical approaches and technological architectures. In addition to identifying trends and gaps in the literature, this study provides insights to inform the development of new educational tools for programming instruction.
@article{arxiv.2509.08857,
title = {A Systematic Mapping Study on Chatbots in Programming Education},
author = {Marcelino Garcia and Renato Garcia and Arthur Parizotto and Andre Mendes and Pedro Valle and Ricardo Vilela and Renato Balancieri and Williamson Silva},
journal= {arXiv preprint arXiv:2509.08857},
year = {2025}
}