Generative AI Supports Path Exploration for Deep Learning in Middle School Science

Abstract

The Compulsory Education Curriculum Plan (2022 Edition) emphasizes the integration of scientific deep learning and digitalization, and generative artificial intelligence injects new momentum into classroom reform. At present, junior high school science teaching faces bottlenecks such as limited experimental conditions, shallow cognition, and lagging feedback, which restricts the development of students' higher-order thinking. In this study, a three-path model of "phenomenon visualization-problem ladder-feedback precision" was constructed, and a scientific inquiry task template was developed based on AI tools, and a controlled experiment was carried out in the eighth grade "Material Change" unit (AI support in the experimental class vs. traditional teaching in the control class). Empirical data show that this path significantly improves students' complex problem-solving rate, experimental design innovation and evidence correlation ability, and promotes the implementation of scientific literacy-based goals. In the future, we will deepen the dual-track integration of AI and physical experiments, build a regional shared task library, and provide a universal paradigm for the digital transformation of science education.