人机协同视域下生成式人工智能辅助高中英语词汇选择题编写研究

A Study on Generative AI-Assisted Senior High School English Vocabulary Multiple-Choice Item Generation from the Perspective of Human-AI Collaboration

摘要: 在生成式人工智能技术快速发展的背景下，AI为语言测试开发提供了新的可能，但现有研究对教师如何运用AI进行人机协同命题的实践过程关注不足。本研究以DeepSeek为辅助工具，探索编写高中英语词汇选择题的可行路径与协作模式。研究发现，DeepSeek在试题格式规范、教材主题契合和语言表达方面具有较好的辅助价值，但AI生成试题仍存在考查内容覆盖不均衡、语境真实性与支持程度不足、正确答案唯一性和干扰项有效性不稳定等问题，需经人工审查与修改方可达到使用标准。研究建议，教师在使用AI辅助命题时应坚持教师主导、人机协同的原则，以提高英语词汇测试开发的效率与质量。

Abstract: Against the backdrop of rapid advances in generative artificial intelligence, AI has opened up new possibilities for language test development. However, existing research has paid insufficient attention to how teachers can engage in human-AI collaborative item generation in practice. This study employs DeepSeek as an assisting tool to explore feasible pathways and collaborative models to develop multiple-choice vocabulary test items for senior high school English. The findings reveal that DeepSeek demonstrates considerable value in terms of item formatting consistency, thematic alignment with textbook content, and naturalness of language expression. Nevertheless, AI-generated items still exhibit limitations, including uneven coverage of target vocabulary, inadequate contextual authenticity and support, and instability in answer uniqueness and distractor effectiveness, all of which necessitate human review and revision before the items can meet quality standards. The study recommends that teachers adhere to the principle of teacher leadership and human-AI collaboration when utilizing AI for item development, so as to enhance both the efficiency and quality of English vocabulary test development.