I. Introduction

Introduces the new opportunities and challenges brought by artificial intelligence technology in the development of English listening teaching materials. It reviews the application research of artificial intelligence in education both domestically and internationally, and clarifies that this study aims to construct intelligent agents to generate high-quality, personalized English listening teaching materials in accordance with the CET-4 standards and question types. The goal is to provide new ideas for English teachers, help learners improve their listening skills, and promote the intelligent development of English education.

II. Research Methods

This chapter introduces the application of intelligent agent tools in the development of English listening teaching materials, as well as the methods for writing and designing prompts. The Kimi+ Prompt Expert and the "Hot Topics Discourse" Agent are two core tools, used to assist in constructing efficient prompts and providing popular text resources, respectively. In terms of prompt writing and design, the structured approach of the AI Teacher Assistant is referenced, with three prompt frameworks designed: title generation, listening comprehension question generation, and vocabulary-to-learn extraction. The title generation prompt framework requires generating Chinese titles, introductions, English titles, and reading outlines based on English articles, with requirements for word count and content of the Chinese titles and introductions. The listening comprehension question generation prompt framework requires generating five questions that meet CET-4 standards based on English articles, specifying the proportion of question types, difficulty distribution, and answer distribution, while proposing detailed question generation protocols and joint verification mechanisms. The vocabulary-to-learn extraction prompt framework requires extracting vocabulary that meets different difficulty requirements from the input text and completing information such as Chinese translations, phonetic transcriptions, definitions, and example sentences. The entire prompt writing and design process emphasizes the importance of format specifications and output checks.

III. Agent Construction and Application

Introduces the construction and application of the English listening practice agent. Based on the Kimi+ Prompt Expert system, the agent can generate listening practice materials that meet specific requirements from English articles, including titles, introductions, listening comprehension questions, and vocabulary to learn. The agent possesses language analysis, pedagogical knowledge, question type design, and vocabulary teaching capabilities, with the goal of generating listening practice questions and vocabulary learning materials that comply with CET-4 standards. The generated listening practice questions must follow CET-4 standards and question type specifications, with Chinese titles not exceeding 20 characters, introductions consisting of 1-3 sentences, listening questions formatted according to requirements, and vocabulary-to-learn output formats conforming to templates. The agent's workflow includes analyzing English articles to extract key information for generating titles, introductions, and outlines; designing five listening comprehension questions; and extracting vocabulary to learn from the articles to create vocabulary learning materials. In the example, the agent quickly generated a Chinese title, introduction, English title, reading outline, and five listening comprehension questions with answers based on an adapted English listening text, demonstrating the practical application effectiveness of the agent.

IV. Research Results and Discussion

Discusses the research results and discussion regarding the construction of intelligent agents for developing English listening teaching materials. The advantages of the generated materials include clear structure, standardized format, and reasonable question type proportions, but there are limitations such as rough reading outlines, low difficulty of listening comprehension questions, and simplistic vocabulary lists. The listening comprehension questions effectively cover the main content of the articles but need improvements in the difficulty of detail-based questions, inference questions, vocabulary selection for main idea questions, and the design of distractors. The vocabulary-to-learn output format meets template requirements, but enhancements are needed in vocabulary sources, difficulty levels, example sentences and usage, and vocabulary exercises. The agent's performance is stable and capable of generating the required teaching materials, but issues exist, including insufficient refinement of reading outlines, unreasonable difficulty gradients in listening questions, and inadequate richness of vocabulary content. Future optimization directions include refining reading outlines, optimizing listening question design, enriching vocabulary content, and improving text adaptability.

V. Conclusion

Emphasizes the importance of English listening skills in the context of globalization and informatization and summarizes the advantages of the English listening practice agent developed in this study. The agent can follow exam standards such as CET-4, generate high-quality listening questions, and provide answer explanations and vocabulary summaries, which help enhance learners' language abilities. Personalized design enables the agent to provide suitable materials based on learners' levels, improving learning efficiency. At the same time, it points out the agent's shortcomings in areas such as reading outline refinement, question difficulty distribution, and vocabulary difficulty levels. It also looks forward to future research directions, including the optimization and improvement of the agent and its potential applications in other language learning fields, to promote the intelligent development of language education.