从片场到课堂：多模态AI赋能的影视韩语ESP教学路径构建——基于DeepSeek行业语料库的语感生成实验

From Studio to Classroom: Building an ESP Teaching Pathway for Film and Television with Multimodal AI Empowerment—A Language Sense Generation Experiment Based on DeepSeek Industry Corpus

摘要: 本研究针对影视类专业韩语教学中的专业认知断层问题，构建多模态AI赋能的ESP教学路径。通过五阶渐进式语料测试(含片场指令、灯光实训、剪辑对话等场景)，验证“教师–学生–AI”三元协同模式的有效性：AI语料库(DeepSeek)精准解析术语并预警语义偏移，使专业词汇掌握率达90%+；真实场景任务设计显著改善相似术语概念混淆，提升提问积极性。实践证明，该方法将教学重心转向高阶思维培养，但需应对AI微语气偏差等挑战。未来需进一步开发专业外语协同课程，推动跨学科师资整合，为艺术院校外语教学探索更多适应性路径。

Abstract: This study addresses the issue of professional cognitive gaps in Korean language teaching for film and television majors and constructs an ESP teaching pathway empowered by multimodal AI. Through a fifth-order progressive corpus test (including set instructions, lighting training, edited dialogues, etc.), the effectiveness of the “teacher-student-AI” ternary collaboration model is verified: the AI corpus (DeepSeek) accurately parses terms and alerts semantic shifts over 90%; real-world task design significantly improves conceptual confusion and enhances question enthusiasm. Practical experience has shown that this method shifts the focus of teaching to the cultivation of advanced thinking, but it also needs to address challenges such as AI microtonal deviations. The future requires further development of professional foreign language collaborative courses, promoting the integration of interdisciplinary teachers, and exploring more adaptive pathways for foreign language teaching at art schools.