<![CDATA[To address persistent issues in junior high chemistry teaching of Composition and Structure of Substances—including fragmented disciplinary perspectives, highly abstract micro-level concepts, and underdeveloped macro-submicro-symbolic triple representation thinking—this study constructs a four-stage interdisciplinary practical teaching model integrated with generative artificial intelligence (AI) and virtual reality (VR). The teaching practice was implemented in a natural ninth-grade class (N=40) in Zhengzhou. To ensure in-depth analysis of the massive multimodal data generated by AI and VR interactions, purposive sampling was employed to select five representative students as a focus group. Following a single-group pretest-posttest design, data from this focus group were evaluated across four dimensions to systematically measure learning outcomes: macro cognition via concept maps, micro cognition via knowledge cards, model deepening through AI text error correction, and transfer competence and AI literacy via AI chat record coding. Post-intervention data revealed significant improvements across all dimensions. In macro cognition, average concept map nodes increased by 108.8% (from 6.8 to 14.2). For micro cognition, accuracy scores of micro-level tasks surged by 170.0% (from 1.0 to 2.7). Furthermore, the error correction rate reached 88.9% during model refinement, while critical probing questions accounted for 45.83% of total chat records, indicating enhanced transfer and AI literacy. These findings confirm that the generative AI-empowered interdisciplinary model effectively broadens students’ cognitive frameworks, lowers learning barriers of abstract chemical knowledge, consolidates mastery of model-based symbolic representation, and fosters preliminary AI critical literacy. It provides a practical, replicable reference for advancing digital interdisciplinary chemistry teaching at the junior high level.]]>
Copyrights © 2026
