Please use this identifier to cite or link to this item: http://repositorio.ufla.br/jspui/handle/1/59853
Title: ChatGPT como recurso de ensino de língua inglesa : análise das potencialidades, limitações e interações linguísticas com a máquina
Other Titles: Chatgpt as a resource for english language teaching: analysis of potentials, limitations, and linguistic interactions with the machine
Authors: Almeida, Patrícia Vasconcelos
Almeida, Patrícia Vasconcelos
Vieira, Mauricéia Silva de Paula
Brandão, Ana Carolina de Laurentiis
Keywords: ChatGPT
Inteligência artificial generativa
Inteligência artificial
Ensino de língua inglesa
CALL
Tecnologia educacional
Generative artificial intelligence
Artificial intelligence
English language teaching
Educational technology
Issue Date: 13-Mar-2025
Publisher: Universidade Federal de Lavras
Citation: MAGALHÃES, Roberto. ChatGPT como recurso de ensino de língua inglesa : análise das potencialidades, limitações e interações linguísticas com a máquina. 144 p. Dissertação (Mestrado em Letras) - Universidade Federal de Lavras, Lavras, 2025.
Abstract: The increasing integration of technologies in education has transformed foreign language teaching practices. In this research, ChatGPT, an Artificial Intelligence (AI) tool, was explored as a pedagogical resource in English language (EL) teaching. The study aims to identify participants' perceptions of the potentialities and limitations of ChatGPT as a tool for EL teaching; analyze participants' perceptions of linguistic interactions with this AI; and analyze participants' perceptions of the process of creating EL teaching activities using this technology. To support the theoretical discussions, the research is based on studies by Warschauer (1996), Levy (1997), Egbert (2005), Levy and Stockwell (2006), Hubbard and Levy (2006), Garret (2009), Hubbard (2009), and Chapelle (2010) to discuss computer-assisted language learning (CALL). To understand the use of large language models like ChatGPT, works by Alshahrani (2023), Birenbaum (2023), and Rudolph, Tan and Tan (2023), which explore the applications of these models in education, were used. The adopted methodology included an exploratory- descriptive approach, with data collection through online questionnaires applied to undergraduate students of an English Language and Literature teaching program at a federal university. The research adopted a mixed-methods approach from a complementary perspective, considering that descriptive statistical analysis of quantitative data was used to identify general trends, while content analysis of qualitative data, based on Bardin (2010), allowed for a deeper understanding of participants' perceptions. The results indicated that participants recognized the speed, practicality, and creativity of ChatGPT in creating activities, as well as its potential to save time, generate ideas, and offer creative support for pedagogical planning. However, they also pointed out limitations such as the inaccuracy of some responses, lack of reliability of information, and the need for well-crafted prompts for satisfactory results. Additionally, participants expressed interest in continuing to use ChatGPT, including for other purposes, highlighting the importance of teacher mediation to adapt and review the generated content, ensuring its suitability for the educational context. The study concludes that ChatGPT presents significant potential as a support tool for EL teaching, offering agility, creativity, and accessibility in activity creation, but its use requires a critical approach, technical knowledge for prompt formulation, and teacher mediation to ensure the quality and adequacy of the generated content. The research reinforces the need for continuous teacher training for the effective use of AI technologies in the educational context, contributing to the improvement of pedagogical practices and paving the way for future investigations into the impact of these tools on language teaching.
URI: http://repositorio.ufla.br/jspui/handle/1/59853
Appears in Collections:BU - Teses e Dissertações



This item is licensed under a Creative Commons License Creative Commons