"Escolho não corrigir": estratégias de professoras de língua portuguesa no contexto de avaliação textual na era da inteligência artificial

Abstract

Considering the possibility that high school students use digital technologies, such as artificial intelligence (AI), to produce their school assignments partially or entirely, this research aimed to describe and analyze public school teachers’ perceptions of the possible impacts of the popularization of ChatGPT on their students’ text production within the context of teaching and learning Portuguese. To achieve this, a printed questionnaire divided into five parts was administered to a number of teachers from a public school in Campo Grande (MS). One part of the questionnaire involved a simulation of text correction, while the others included questions addressing the proposed theme. The data obtained were analyzed according to a qualitative and interpretative research paradigm, also following theoretical concepts from Antunes (2006) on text evaluation, the typology of essay corrections discussed by Ruiz (2001), critical considerations on AI by various authors (Buzato, 2023; Chomsky; Roberts; Watumull, 2023; Weizenbaum, 1976), and its application in the educational context (Rodrigues; Biondo, 2024; Carmo; Carmo; Melo, 2022). In conclusion, it was possible to show, based on the teachers’ responses, that they perceived signs of interference from the popularization of ChatGPT in their students' textual productions and that they hold a negative view of AI use in the context of school text production.