Letramento digital em Inteligência Artificial: conhecimentos e percepções de Pós-Graduandos em Educação sobre IA na pesquisa científica acadêmica

Abstract

This research analyzed the intersection between Artificial Intelligence (AI), Education, and Academic Scientific Research from a critical and ethical perspective, identifying the profile of Digital Literacy in AI (LDIA) among postgraduate students in Education at the Federal University of Ceará (UFC). It was developed using a mixed approach, exploratory bias, with an online questionnaire applied to 148 subjects in May 2025. The results show a sample composed of 60 master's students and 88 doctoral students, with an average age of 38.1 years, predominantly female (62.6%), licensed (82%), with first contact with AI between 2022-23 (62.3%). The vast majority (94.6%) have already used AI tools, the most cited being: ChatGPT (34.11%), Gemini (10.47%), DeepSeek (6.20%), and Copilot (5.43%). 9.4% fear the use of AI for stealing ideas, while 91.2% believe in systemic impacts, especially in education, in a positive way (58.1%). AI is mostly seen as a virtual assistant (50.0%) and artificial neural network (29.8%), while 69.6% believe that AI is a technology that combines software and hardware in the execution of tasks such as planning and actions that require logical reasoning and creativity. In the area of research, subjects believe that it can contribute by translating (85.8%) and finding scientific literature (68.2%), reviewing writing (67.6%), illustrating data (79.8%), generating slides (63.5%), and creating educational resources (76.4%). On the other hand, 22.3% believe that using AI constitutes plagiarism. 53.4% have never participated in an event or training in the area, with 83.8% expressing interest in doing so. 52.7% intend to use AI in their research, describing themselves as curious but with doubts (48.1%), with a demand for literacy. The hypothesis, which was confirmed, was that AI contributes to research, as long as the researcher's intellectual autonomy is maintained in the process. It was concluded that graduate students use AI, have a high theoretical awareness of it, but are unaware of specific tools, revealing themselves to be optimistic but cautious. This data reinforces the need for training actions on critical and conscious use, that is, LDIA practices, highlighting the balance between technological innovation and human protagonism, based on ethics and responsibility.