Academic literacies and Artificial Intelligence: analyzing the simulation of scientific article comprehension through ChatGPT

Objective: Artificial intelligence (AI) has permeated various contexts, including academia, sparking debates about its potential and limitations in this context. This study aims to analyze the main characteristics of responses generated by the ChatGPT tool when prompted with questions about the composition of academic articles. Methods: Adopting a descriptive and exploratory approach, this study focuses on interactions with ChatGPT. Five sequential prompts, presented in both Brazilian Portuguese and English, were designed to progressively increase in linguistic complexity. Results: The analysis revealed that ChatGPT exhibits proficiency in synthesizing information and producing responses that align with academic conventions, highlighting its utility as a research support tool. However, the responses predominantly reflected a skills-based model, emphasizing technical aspects and the formal structure of academic writing such as organizational clarity and methodological rigor-while neglecting the social and identity dimensions of academic authorship. Conclusions: While the guidance provided by ChatGPT is valuable for understanding the technical aspects of academic article composition, it is essential for users to cultivate a critical awareness of writing practices. Recognizing the role of academic literacies is crucial for interpreting and constructing knowledge within the academic landscape.