Inteligência artificial generativa no ensino de física: modelagem da lei do resfriamento de Newton com Gemini e ChatGPT

Abstract

This research investigates the potential of Generative Artificial Intelligence tools, specifically Gemini and ChatGPT, as a didactic resource for the mathematical modeling and graphical representation of Newton's Law of Cooling in high school. The work aimed to support the development of a didactic sequence based on Inquiry-Based Learning. The methodology adopted consisted of a comparative simulation, using a single, standardized prompt in both Generative Artificial Intelligences, providing fixed boundary conditions (initial and ambient temperature) and cooling coefficients (k) for the materials steel, copper, and aluminum. The execution compared the accuracy of the generated equations and the fidelity of the individual and overlay graphs, in addition to evaluating the usability and functionality of interactive parameter manipulation. The results demonstrated that both tools achieved mathematical accuracy in generating the formula for Newton's Law of Cooling. However, Gemini exhibited superiority in graphical fidelity and stability of use. Furthermore, the graphs generated by Gemini were more suitable, presenting different colors for each material and including a clear line representing ambient temperature, all without the operational restrictions observed in ChatGPT. It is concluded that Generative Artificial Intelligence is a powerful virtual laboratory, whose use is enhanced by active methodologies, guaranteeing student protagonism. The resulting Didactic Sequence integrates Gemini as a priority tool for simulation, this Didactic Sequence being of fundamental importance for guiding pedagogical practice, and the teacher can apply it in the classroom for high school students, promoting investigative learning and digital culture, which is directly aligned with the General Competencies of the National Common Curricular Base.