生成式人工智能辅助大学英语阅读测试文本开发研究

过君琰

doi:10.70693/202609285674

文献检索 | 登录 | 注册

开放获取　·　同行评审　·　学术期刊

人文与社会科学学刊维普检索国际期刊

《人文与社会科学学刊》是由香港 Wisvora Publishing 出版、面向全球发行的国际中文学术期刊。本刊聚焦人文社科领域的理论创新与实践应用，深耕学科核心议题与前沿研究，倡导跨学科、跨文化的多元研究视角，打通学术研究与社会实践的沟通壁垒，推动学术知识的传播、共享与转化，助力全球华语学术界交流与发展。

在线投稿模板下载最新一期

ISSN 3079-4552 发行周期：月刊发文量：973 篇

公告 《科学与工程研究》创刊征稿覆盖理工科各学科方向 新刊《科学与工程研究》为双月刊，论文将被维普数据库收录，面费优惠价 880 元 / 篇，欢迎广大科研工作者踊跃投稿。详情

在线投稿下载模版过刊浏览

切换期刊

期刊栏目

加入编委

期刊订阅

添加您的邮件地址以接收即将发行期刊数据：

Open Access Article

PDF

生成式人工智能辅助大学英语阅读测试文本开发研究

Developing GenAI-assisted College English Reading Test Texts

作者：过君琰

单位：无锡学院

基金项目：无锡学院2024年“公共基础课程教学质量提升”专项课题“生成式人工智能辅助大学英语阅读测试资源库建设研究”(2024JGKT011)

卷期：2026年10月-2卷10期

通讯作者：过君琰

DOI：10.70693/202609285674

页码：162-167

发布时间：2026-09-29

总浏览量：187

摘要

生成式人工智能的快速发展为外语测试资源的开发提供了新的技术路径。阅读是大学英语学习与语言能力评价的重要组成部分，也是大学英语四、六级考试的核心测试内容之一。因此，探索生成式人工智能辅助阅读测试资源开发具有重要的教学和测评价值。本研究考察生成式人工智能（ChatGPT）在大学英语四、六级阅读测试文本开发中的应用潜力，重点展开两类生成任务：一是以四、六级真题为参照，针对真题阅读的三种题型，分别生成相应层级的阅读测试文本；二是以大学英语教材课文为基础，将其改编成适用于四、六级三种题型的阅读测试文本。本研究从文本难度和词汇水平两个维度，将两类任务生成的文本与真题进行比较。结果表明，生成式人工智能能够基本按照目标层级和题型要求生成和改编文本，其文本难度和词汇水平与真题总体较为接近。本研究为生成式人工智能赋能外语阅读测试资源开发提供了实践启示。

关键词

生成式人工智能；大学英语；阅读测试；文本开发

Abstract

The rapid advancement of Generative Artificial Intelligence (GenAI) has offered a new technological approach for the development of resources for foreign language assessment. Reading is an essential component of college English learning and language proficiency assessment and constitutes a major part of the College English Test (CET). Therefore, exploring the use of GenAI in developing reading test resources has significant pedagogical and assessment value. This study investigates the potential of GenAI (ChatGPT) for developing reading test texts for CET-4 and CET-6 through two tasks. First, using authentic CET-4 and CET-6 reading texts as benchmarks, GenAI was employed to generate texts at the corresponding proficiency levels for three reading task types. Second, a text from a college English textbook was adapted into CET-4- and CET-6-level reading texts for the same three task types. The two sets of GenAI-generated texts were then compared with authentic CET texts in terms of text difficulty and vocabulary level. The results showed that GenAI was generally capable of generating and adapting reading texts according to specified proficiency levels and task requirements. The generated texts were broadly comparable to authentic CET texts in text difficulty and vocabulary level. This study offers practical implications for GenAI-powered foreign language reading test development.

Keywords

Generative Artificial Intelligence; College English; Reading test; text development

下载PDF

引用本文

过君琰. 生成式人工智能辅助大学英语阅读测试文本开发研究[J]. 人文与社会科学学刊. 2026, 2 (10): 162-167. DOI: 10.70693/202609285674.

APA引用

过君琰. (2026). 生成式人工智能辅助大学英语阅读测试文本开发研究. 人文与社会科学学刊, 2 (10), 162-167. https://doi.org/10.70693/202609285674

参考文献

[1] CHUANG P L, YAN X. (2025). Language assessment in the era of generative artificial intelligence: Opportunities, challenges, and future directions[J]. System, 2025, 134. https://doi.org/10.1016/j.system.2025.103846
[2] 刘建达. 人工智能时代的语言测评：机遇与挑战[J]. 现代外语, 2024, 47(6), 859–869.
[3] ABUDALFA S I, BARROT J. S. Generative artificial intelligence for automated writing evaluation: A systematic review of trends, efficacy, and challenges[J]. Assessing Writing, 2026, 68, 101041. https://doi.org/10.1016/j.asw.2026.101041
[4] ANBREEN T, ÖZTURAN T, SHRESTHA P, et al. Human vs. generative artificial intelligence in writing assessment: investigating feedback alignment, score validity, and teacher agency[J]. Educational Assessment, Evaluation and Accountability, 2026, https://doi.org/10.1007/s11092-026-09486-z
[5] GOH C C M, ARYADOUST V. Developing and assessing second language listening and speaking: Does AI make it better?[J] Annual Review of Applied Linguistics, 2025, 45, 179–199. https://doi.org/10.1017/S0267190525100111
[6] GUO J. Exploring GenAI-powered listening test development[J]. Languages, 2026, 11(1). https://doi.org/10.3390/languages11010017
[7] LIU X J, WANG J, ZOU B. Evaluating an AI speaking assessment tool: Score accuracy, perceived validity, and oral peer feedback as feedback enhancement[J]. Journal of English for Academic Purposes, 2025, 75, 101505. https://doi.org/10.1016/j.jeap.2025.101505
[8] 张海云. 大学英语阅读诊断反馈报告应用及评价——以G-DINA模型为例[J]. 外语电化教学, 2022, 39-47,117.
[9] 中华人民共和国教育部高等学校大学外语教学指导委员会. 大学英语教学指南（2020版）[M]. 北京：高等教育出版社, 2020: 12-16.
[10] 中华人民共和国教育部、国家语言文字工作委员会. 中国英语能力等级量表[S]. 上海：上海外语教育出版社, 2024: 10.
[11] 全国大学英语四、六级考试委员会. 全国大学英语四、六级考试大纲（2016年修订版）[EB/OL]. 2016: 2-8.
[12] LIN Z, CHEN H. Investigating the capability of ChatGPT for generating multiple-choice reading comprehension items[J]. System, 2024, 123, 103344. https://doi.org/10.1016/j.system.2024.103344
[13] WEN Z, CHU S K W. Using generative AI for reading question creation based on PIRLS 2011 framework[J]. Cogent Education, 2025, 12(1), 2458653. https://doi.org/10.1080/2331186X.2025.2458653
[14] 罗娟, 黄骞. 生成式人工智能与人工命题的差异研究：以高考英语阅读理解命题为例[J]. 基础外语教育, 2025, 27(4), 97-106,112.
[15] 陈大建, 胡杰辉. 基于大语言模型的自动化命题研究——以英语阅读理解试题为例[J]. 外语教学, 2025, 46(2), 40–47.
[16] NÁBĚLKOVÁ J, KOPECKÝ K, PLISCHKE J, et al. Criterion-based assessment of reading literacy with AI[J]. Frontiers in Education, 2026, 11. https://doi.org/10.3389/feduc.2026.1812149
[17] 吴晓真, 陈进. 全新版大学进阶英语综合教程（第二版）第三册[M]. 上海：上海外语教育出版社, 2024: 14-16.
[18] 金檀, 陆小飞等. (2025). 英文分级指难针[EB/OL]. 一针三库智能教研平台(languagedata.net).
[19] NATION I S P. How large a vocabulary is needed for reading and listening?[J] The Canadian Modern Language Review, 2006, 63(1), 59–82.