Uso do ChatGPT como tutor para a configuração de ambientes de simulação de redes de computadores

Abstract

This work addresses the configuration of development environments designed for the analysis of computer networks, encompassing the preparation of virtual machines, software installation, and gathering of materials for the study of computer network simulation using the Generative Pre-Trained Transformer (GPT) algorithm in version 3.5, utilized as the foundation of the ChatGPT tool. The study analyzes how ChatGPT significantly aids in the configuration of environments and the study of computer network simulations. In the academic context, many students face learning disorders such as dyslexia, dyscalculia, autism, or attention disorders (deficit or hyperactivity). Especially during the pandemic period, when students had to adapt to remote learning, these students had to find ways to study independently. At this moment, the absence of a present tutor generates significant difficulties in exploring new subjects and areas of knowledge. Hence, there arises the need to explore new ways of acquiring knowledge, and the idea of evaluating the potential of ChatGPT as an intelligent assistant. Particularly, its use by students initiating their studies in computer networks. This work seeks to assess the ChatGPT tool in terms of responses, providing detailed guidance on procedures for configuring network simulation environments, explanations of terminology, and content suggestions for study. Using two distinct forms of requesting help, one where a single chat would be opened, and another where a new chat is opened for each different topic, we defined different topics as follows: virtual machine installation, virtual machine configuration, Core Emulator installation, Wireshark installation, definitions of technical terms, and analysis of logs generated by Wireshark. As these are distinct topics with extensive configurations, it became clear that seeking assistance from ChatGPT by opening new chats for each topic demonstrates greater effectiveness compared to attempting to seek help directly within a single chat. It is crucial to remember that the tool has limitations, such as dependency on internet connectivity and the possibility of providing imprecise answers. Thus, while the artificial intelligence represented by ChatGPT is a valuable resource, the tool sometimes returned inconclusive or incomplete answers, showing that it is not capable of entirely replacing the human expertise of an experienced tutor. It remains a complementary tool, highly useful in the context of studying computer network simulation, as well as in the configuration of environments and the study of routing protocols.