Transformation of Students' Scientific Reasoning Through Artificial Intelligence-Based Media: A Systematic Review

Abstract

This study aims to examine the transformation of students' scientific reasoning through the use of Artificial Intelligence (AI)-based media in science education. The method employed is a Systematic Literature Review (SLR), with a comprehensive literature search conducted in the Scopus database. This study reviews various articles that discuss the application of AI in science education, particularly focusing on its impact on students' scientific reasoning. The findings show that AI significantly contributes to enhancing students' abilities to analyze data, identify patterns, and understand complex scientific concepts. AI technologies, such as Graph Neural Networks (GNN) and generative AI models (VAE and GAN), enable students to conduct virtual experiments and simulations that accelerate their understanding of science content. Additionally, AI supports adaptive learning tailored to students' individual needs, allowing them to learn at their own pace. In conclusion, integrating AI into science education not only accelerates the learning process but also deepens students' scientific reasoning, equipping them with critical thinking skills necessary to tackle future challenges. The implications of this study suggest that AI can be an effective pedagogical tool in enriching students' learning experiences, and further exploration of its use in various educational contexts is essential.