The application of Generative Artifi cial Intelligence in the fi eld of education is gradually deepening, demonstrating unique technical advantages and practical potential in high school geography teaching.As a typical representative of such tools, DeepSeek has shown outstanding performance across multiple technical dimensions and is regarded as a signifi cant force in promoting the intelligent transformation of teaching. However, due to inherent limitations of the model in terms of data quality, comprehension of subject, and teaching adaptability, its generated content often fails to fully align with curriculum objectives and practical teaching needs.This necessitates teachers to exercise professional judgment in carefully identifying and correcting deviations in the generated results. This paper takes DeepSeek as the research subject and incorporating a specific case study from high school geography“Locational Factors of Service Industries and Their Changes”.Then, it explores the human-machine collaborative correction paths applicable to the geography discipline, with the aim of providing references for the standardization of generative artifi cial intelligence's educational application and the professional development of teachers.