Inteligência artificial e “geração” de texto dissertativo-argumentativo: limites da escrita por máquinas.

Abstract

In the age of Artificial Intelligence, a common topic from a writing perspective is the possibility of using machines to generate texts. However, a recurring question arises regarding the "generation" process and the limits established in such a situation, particularly in the context of the ENEM (National High School Exam) essay. Therefore, the objective of this article is to analyze the development of ideas for a ChatGPT-generated dissertation-argumentative text compared to a 100-point dissertation-argumentative text produced by an ENEM candidate. As a theoretical basis, we draw in-depth on Koch and Elias (2016), Hirota (2023), Vicari et al. (2023), Fernandes (2024), Cavalcante and Lemos (2023), among others. Regarding the methodology, this is an analytical and interpretative study (Severino, 2013), because it involves comparative analysis; and a documentary study (Severino, 2013), as it uses a text taken from the ENEM 2024 Participant's Handbook (Brazil, 2024). The results prove that AI is a complementary alternative and that the texts should not be copies of the results obtained in the chatbot, since there is a compromise in reading and writing skills, in addition to the disruption of authorship and the incentive to the formation of student-writers (Fernandes, 2024).