GENERATIVE ARTIFICIAL INTELLIGENCE IN THE TEACHING OF DIFFERENTIAL AND INTEGRAL CALCULUS: A COMPARATIVE ANALYSIS OF DIFFERENT AI MODELS

Abstract

This study analyzes the potential of different generative artificial intelligence tools as resources to support the teaching and learning of Differential and Integral Calculus in Brazilian higher education. Motivated by the growing adoption of these technologies in educational settings and the persistent challenges associated with high failure rates in calculus courses, the research aimed to compare the quality of responses generated by five AI systems (Gemini Free, Gemini Pro, ChatGPT Free, Perplexity AI Pro, and Grok AI Free) when solving the same set of problems involving limits, derivatives, and integrals. The study adopted a qualitative, descriptive, and comparative approach, using a standardized prompt that required detailed step-by-step solutions, conceptual explanations, identification of common mistakes, and verification of results. The findings revealed that all AI systems correctly solved the proposed problems; however, differences were observed regarding organization, clarity, mathematical rigor, notation quality, and pedagogical depth. Gemini Pro and Perplexity AI Pro demonstrated the highest level of consistency and instructional quality, while ChatGPT Free and Gemini Free also proved to be effective tools for supporting students' learning. The study concludes that generative artificial intelligence has considerable potential to promote personalized learning, provide immediate feedback, and foster students' autonomy, provided that it is used ethically, critically, and in conjunction with effective pedagogical guidance.