从“作业”到“智业”：Deep Seek 赋能小学数学教师设计分层任务的路径研究

Abstract

In the context of the "Double Reduction" policy, aiming at the dilemma of homogeneous primary school mathematics homework, this study explores the innovative paths through which the DeepSeek intelligent platform empowers teachers to design hierarchical homework. From the perspective of the collaborative design between teachers and generative AI, this research proposes a "Four-Stage Reconstruction Model" for hierarchical homework in primary school mathematics. Through the qualitative case analysis method, taking the unit of "Fractional Operations" as an example, it systematically demonstrates how teachers can complete the whole process of "needs diagnosis - hierarchical adaptation - intelligent generation - teaching iteration" with the help of DeepSeek. The study finds that the structured prompting strategy can transform teachers' pedagogical knowledge into design instructions recognizable by AI, generating a three-dimensional task system that includes the basic consolidation layer, the transfer and application layer, and the expansion and exploration layer, providing a practical paradigm for the transformation of homework under the background of the "Double Reduction" policy.