Exploring the Application of Generative AI in University English Teaching Reform

The technology of generative artificial intelligence (GAI) can be traced back to the 1950s, but it was not until the early 21st century, marked by the emergence of deep learning, that GAI entered a new stage of development (Dhamani & Engler 2024). Especially after OpenAI launched the ChatGPT chatbot based on large language models in 2022, people have genuinely felt the revolutionary impact of GAI on various aspects of social life, particularly in work and education (Yang Zongkai et al. 2023). The Association for Educational Technology in Higher Education in the United States, in its “2024 Horizon Report: Teaching and Learning Edition” published in May 2024, for the first time separated AI, represented by ChatGPT, from other information technologies, dedicating a section to discuss in detail the impact of AI on higher education in areas such as teaching content, teaching models, teaching evaluation, classroom management, and student mental health (Pelletier et al. 2024).

So far, research on AI development and application has mostly focused on the technical level, neglecting the role of the human subject and leading factors in technology development and application (Xu 2019). However, it is worth noting that some studies in recent years have begun to focus on the relationship between humans and technology in technology development and application. For example, some research points out that, from the perspective of the purpose of technology development and application, it should be clear that technology should ultimately serve humanity, and the development of AI should reflect the depth unique to human intelligence, enhancing rather than replacing human capabilities (Xu 2019). Other studies have proposed that the application paradigm of information technology should shift from the past “AI-dominant” model, where learners are receivers, to an “AI-empowered” model, where learners are leaders, promoting the integration of human intelligence and artificial intelligence (Ouyang & Jiao 2021: 4). This article starts from the human-centered shift in AI technology application (Li Xue, Gu Xiaole 2024; Lü Guangzu, Shi Miao 2024) and, based on the era’s requirements for university English teaching reform and the characteristics of foreign language education, discusses the practical path of GAI empowering university English teaching reform using the author’s practical experience in the GAI application in the “General Academic English Writing” course as an example.

According to the “University English Teaching Guidelines (2020 Edition)” (hereinafter referred to as “Guidelines”) (Ministry of Education Higher Education Foreign Language Teaching Guidance Committee 2020), the “General Academic English Writing” course falls under the category of academic English courses within the specialized English courses. This course is aimed at undergraduate students in their first and second years who have reached a basic level of university English and have a clear need for general academic English application but lack the ability to conduct high-level professional research in English. The course aims to help students transition from general English to specialized English course learning. General academic English courses have become relatively mature in the field of university English teaching in China, and in recent years, the offering of this course has been on the rise in various universities, especially in high-level institutions (Yao Pengfei, Wang Yong 2020). At Nanjing University, where the author works, this course has been offered for about 20 years (Wang Wenyu et al. 2018). However, like other courses, with the changes of the times, especially with the increasing popularity of GAI applications, including large language models, and the new requirements continuously put forward by the state for university English teaching, as well as the deepening understanding of English teaching theory and practice, it is necessary to systematically reconstruct and enhance academic English courses in terms of teaching objectives, content, methods, and evaluation (Ralph Taylor 1949/1994). Below, we will select typical cases from the eight aspects of revising the course outline to explore the implementation path of GAI empowering university English course teaching reform.

2.1 Course Outline

The course outline is a guiding document for course construction, and one of the primary tasks in formulating the course outline is to clarify the teaching objectives, which depend on the requirements of the nation and society, the characteristics of the discipline, and the needs of the learners themselves (Ralph Taylor 1949/1994). The “Guidelines” state that university English teaching is an important component of higher education and should “play an important role in implementing the fundamental task of cultivating virtue and fostering talent in higher education,” meeting the needs of national, school, and individual student development. It should also “continuously update concepts and use methods that align with the characteristics of contemporary university students to carry out teaching activities.” To better implement the requirements of the “Guidelines,” the primary task of optimizing the course outline for the “General Academic English Writing” course is to optimize the teaching guiding philosophy and teaching requirements. To this end, we have adopted a two-step approach.

The first step is to use GAI’s literature summarization function to generate the following three categories of teaching guiding principles for the General Academic English Writing course:

(1) Course ideological and political education requirements

(2) Project-Based Language Learning with Chinese Characteristics (PBLL-C) theory and application principles

(3) Information technology application principles represented by GAI

The second step is to guide GAI to generate the course description and teaching objectives based on the above three categories of teaching guiding principles.

Tables 1 to 3 respectively present the summary of the teaching guiding principles generated by GAI, the course overview, and the course teaching objectives. It can be seen that the text can adequately reflect the teaching objectives and requirements we hope to achieve in this course.

In terms of prompt design ideas, this case adopts the strategies of “providing background knowledge” and “step-by-step implementation” (Bozkurt 2024). On one hand, although large language models have a broad knowledge base, their understanding of specific content related to university English course ideological education, PBLL-C teaching principles, etc., is still quite limited. Therefore, they need to carefully study relevant literature with the user’s help to accurately understand social, disciplinary, and student needs before designing course content that better meets the needs of the nation, the times, and the students. On the other hand, clear teaching objectives require guidance from scientifically systematic teaching principles, which in turn need to be based on the analysis and understanding of relevant specialized literature. This process involves multiple steps, and each step may involve multiple sub-steps. Given the working characteristics of large language models, it is necessary to break down the entire task into different sub-tasks for step-by-step implementation to generate accurate and practical teaching objectives using AI.

2.2 Teaching Plan

The teaching plan is a detailed teaching scheme derived from the course outline and is an important text for implementing the course outline. To ensure that the course teaching objectives can be truly realized in teaching, the teaching plan must ensure that the teaching elements such as objectives, content, methods, and evaluations present a linear extension while also reflecting interconnection and cyclical characteristics (Ralph Taylor 1949/1994) at a macro level, and at a micro level, it must help teachers design student-centered teaching activities while closely integrating academic English writing skills with the themes of the textbook units to ensure the coherence, progression, and practicality of the teaching content. To achieve the above goals, we first generated an outline of the semester teaching plan based on the previously generated course principles, teaching objectives, and course descriptions, and then generated weekly teaching plan outlines one by one based on this outline. In the subsequent teaching process, we can also generate implementable weekly teaching plans based on the weekly teaching plan outlines (this article is abbreviated). The relevant GAI prompts and the generated semester teaching plan outline and weekly teaching plan outlines are shown in Tables 4, 5, and 6.

2.3 Teaching Content

One of the most important steps in course construction is selecting teaching content and materials. Existing academic English writing textbooks contain rich content on writing skills and guidance on the writing process, and some textbooks even provide academic English examples for teachers or students to analyze, learn, or imitate. However, in actual teaching, both teachers and students encounter many issues, such as fragmented teaching content and mismatches between skill teaching and writing examples. In such cases, GAI can help teachers generate high-quality teaching content.

For example, the textbook for this course includes teaching content on how to write topic sentences and how to use evidence to support the author’s viewpoint, but it lacks examples related to this teaching content. Even if there are a few examples, they often come from different texts and are not related in content. To achieve better teaching results, we first used GAI to generate ten strategies that relate topic sentences and evidence (see Table 7), and then asked GAI to write an essay of the same topic and format, requiring GAI to apply all ten strategies (see Table 8), and then to explain in detail where these ten strategies were used and how their application helped express the content of the article.

The prompt design strategy used in this case is “role-playing” (White et al. 2023) under the category of “role transformation.” First, GAI generates an output as a teacher, which is the ten writing strategies, then as a student, applies these strategies to write an essay, and finally, as a writer, explains the use of each strategy in the essay. The complete chapter generated with the help of GAI not only helps teachers concentrate the previously fragmented writing skills teaching into one article but also makes the generated essay directly related to the writing task that students are currently engaged in. This teaching method, which combines specific writing tasks with comprehensive learning of writing skills, should help improve teaching effectiveness.

2.4 Teaching Resources

Compared to traditional AI, the greatest characteristic of GAI is its ability to autonomously generate new content based on user instructions, such as text, images, audio, and video (Maslej et al. 2024). In higher education teaching, teachers can comprehensively utilize different GAI tools to develop teaching resources and materials that meet the needs of teaching reform in aspects such as content generation, media processing, format conversion, and result presentation. In the revision of the teaching plan for this course, we added the theme of academic English vocabulary, and to support the teaching of this theme, we needed to produce a short video. The production of short videos or micro-lectures involves creating lecture content and scripts, video recording, and post-production. If using traditional methods to record a 5-10 minute micro-lecture, it generally takes one to two days or even longer to complete, and it requires the joint participation of both the teaching team and the technical team. At the same time, in the production of this “academic English vocabulary” micro-lecture, we also learned that although there are now many apps claiming to generate lecture videos in one step, these products that require no human involvement often have many issues in content, voice, and visuals. To efficiently produce high-quality micro-lectures, we divided the production process into content outline and script generation, speech points and slide file generation (using mindshow.fun), video generation based on slide files, and script voice synthesis (using Microsoft’s Azure platform), and finally used conventional digital video editing software (using the “Jianying” platform) to complete the micro-lecture video synthesis (as shown in Figure 1).

2.5 Teaching Models

In designing teaching reform plans, GAI supports the design and implementation of various teaching models, flexibly applying online, offline, hybrid, and virtual simulation models to create the optimal combination of teaching models, ensuring the effectiveness of teaching activities and the student learning experience, allowing every student to have equal educational opportunities and resources (Akinwalere & Ivanov 2022). For instance, one of the teaching objectives of this course is to help students acquire preliminary research abilities. Since students in this class come from different departments, interdisciplinary collaboration and communication skills are also important components of research abilities. To better organize virtual simulation collaborative research classroom interactions, the teacher used GAI to complete the following steps: (1) generate collaborative research scenarios; (2) assign roles and tasks to students, with GAI playing one role and communicating with students; (3) generate a scenario where a conflict occurs during communication and work together with students to resolve the conflict; (4) help students report research results and answer questions that the audience may raise. The prompts used in each step of organizing this activity are shown in Table 9.

2.6 Teaching Methods

In terms of innovating teaching methods, GAI can play an important role. By providing suggestions for designing teaching activities and supporting teaching resources, GAI can help teachers organize diverse teaching activities such as individual autonomous learning, group cooperative learning, classroom and online interactions, and thematic seminars, providing students with personalized learning paths and resource recommendations (Huang Libo 2022). These teaching activities can not only promote students’ autonomous learning and collaborative communication but also enhance the interactivity and engagement of teaching (Li 2023; Zhu Zhitian et al. 2023).

Since students in this course come from different departments, each student’s learning objectives and content may vary significantly. Therefore, at the beginning of the semester, students are encouraged to construct personal learning plans based on their learning interests and abilities, set specific learning objectives, and achieve their personal learning goals with the feedback and guidance provided by GAI, thereby enhancing their motivation and self-management abilities in learning English courses. Considering the dynamic and phased characteristics of this task, students at the beginning of the semester generate descriptions of their writing levels based on their writing outputs and the teaching requirements of this course using GAI tools, forming preliminary learning plans, and then regularly monitor their language levels and learning inputs using GAI tools during the semester, adjusting learning content as needed, and at the end of the semester, reflecting on the learning process with GAI to form the next learning plan and scheme. Specific steps and suggested prompts are shown in Table 10.

In this teaching case, the design of prompts not only employs strategies such as “step-by-step implementation” and “role-playing” but also incorporates the “data analysis” strategy. Large language models, in addition to excelling in conventional language interaction, also possess certain applications in statistics and corpus linguistics, allowing them to generate scientifically standard inferential conclusions or reasonable suggestions based on the provided data and corpora (Dieruf 2023).

2.7 Teaching Tools

GAI has significant advantages in the development of teaching tools. By providing functions similar to dictionaries, grammar textbooks, and corpora, GAI can help students look up and learn relevant knowledge at any time during the learning process, thus improving learning efficiency. GAI-driven intelligent writing platforms can provide students with real-time writing feedback and suggestions, helping them improve writing quality and efficiency.

For example, in teaching, if a student wants to express “warm” in “warm water” in English, the traditional method would be to consult a Chinese-English dictionary. However, regardless of how comprehensive a physical dictionary is, the explanatory content it can provide is still limited, not to mention the many inconveniences of using a physical dictionary. Now, with GAI, students can ask their questions and get answers at any time (see Table 11).

In this case, the design of prompts not only clarifies the problems to be solved but also employs the strategy of “clarifying output requirements” (Mantri 2023). When using this strategy, users can not only specify requirements for the content of the responses but also for the presentation format, file format, etc., such as requesting the large language model to generate the required content in list format, presentation format, PDF format, chart format, image format, etc.

2.8 Teaching Evaluation

In terms of teaching evaluation, GAI can analyze students’ learning performance and data to provide professional and continuous evaluation and feedback, helping them reflect and improve throughout the learning process; GAI can also help teachers better understand students’ learning situations, adjust teaching strategies, achieve precise teaching, and promote personalized learning (Pelletier et al. 2024).

In this course, providing feedback on students’ writing is an important teaching content, and precise, high-quality writing feedback can effectively promote the improvement of students’ second language proficiency (Su Qi 2024). However, in writing teaching, grading assignments is a very time-consuming and labor-intensive task that is often difficult to gain students’ attention. Most existing GAI tools already possess certain capabilities for evaluating and providing feedback on students’ writing. These tools can evaluate both the overall quality of the text and its specific aspects, providing service to students throughout the entire process and timeframe. Research comparing the evaluation quality of language teachers and large language models has found that both have strengths and weaknesses in different evaluation dimensions, but large language models perform better than language teachers in providing standard-based feedback and implementing process-oriented evaluations (Steiss et al. 2024). In light of this, for the semester paper writing in this course, we designed a dedicated intelligent agent for providing feedback on long English papers (as shown in Figure 2).

This intelligent agent embeds specific requirements for the assignment in terms of theme, content, format, norms, etc. When students use it, they can obtain evaluations and modification suggestions from the preset aspects of “Theme and Evidence,” “Argumentation and Logic,” “Discourse and Structure,” “Originality and Completeness,” “Accuracy and Appropriateness,” and “Norms and Formatting.” This application of the large language model, which combines specific assignment requirements, enhances the accuracy of GAI’s writing feedback and the preset evaluative choices not only leverage GAI’s strengths and improve the quality of writing evaluation and feedback but also have a good promoting effect on students’ learning of writing knowledge, helping students apply the writing knowledge and skills taught in a scattered manner throughout the course to their own paper writing. This GAI-assisted writing training can not only improve the quality of papers but, more importantly, promote the enhancement of students’ writing abilities (Xu Linlin et al. 2024).

This article combines the teaching reform practice of the “General Academic English Writing” course to explore the application paths of GAI in university English course teaching reform from eight aspects: curriculum outline, teaching plan, teaching content, teaching resources, teaching models, teaching methods, teaching tools, and teaching evaluation. Based on the above case analyses and existing research on GAI applications, we believe that in the construction and teaching design of university English courses, reasonable use of GAI tools and scientific design of prompts should focus on five aspects.

First, adhere to a human-centered approach and explore GAI applications based on the needs of teaching reform. The GAI application based on the concept of technological empowerment requires teachers to first start from the teaching reform objectives of the course and the learning needs of students, adhering to a teaching philosophy that emphasizes teacher leadership and student-centeredness, paying attention to the new concepts and requirements proposed by the Party and the state for higher education, especially for university English teaching, ensuring that GAI applications contribute to achieving the predetermined teaching objectives, helping teachers explore new teaching models such as flipped classrooms and project-based learning; guiding GAI to design project tasks and discussion questions, promoting the development of students’ innovative thinking abilities, higher-order thinking abilities, autonomous learning abilities, and collaborative spirit, and supporting students’ comprehensive development (Yang Zongkai et al. 2023; Zhou Hongyu, Chang Shunli 2023).

Second, organically integrate GAI applications into teaching design to enhance the teaching and learning experiences of teachers and students. Teachers should consider GAI characteristics while designing teaching, based on language teaching laws, course teaching objectives, and students’ language levels, learning styles, and interests, organically integrating GAI applications into the teaching process, using GAI to design innovative teaching activities, generating suitable learning materials and guidance for students with different learning styles and needs, stimulating students’ learning interests, creating interactive and personalized learning experiences, and promoting deep learning (Kong Lei 2024; Wen Qiufang 2024; Xu Jiajin, Zhao Chong 2024).

Third, continuously enhance the digital literacy of teachers and students, enriching the basic knowledge and application skills of GAI applications. University English teachers and students should continuously improve their GAI application literacy through active learning and practice, effectively mastering the basic knowledge and skills related to artificial intelligence, prompt design, and GAI tools (Knoth et al. 2024; Wen Qiufang, Liang Maocheng 2024). For example, in the design and application of prompts, adhere to the principles of clarity, specificity, accuracy, and iteration. In addition to the prompt design strategies introduced in this article, such as “step-by-step implementation,” “outline expansion,” “role-playing,” and “clarifying output requirements,” more strategies like “multi-turn prompts” and “reflective prompts” can also be employed (Korzynski et al. 2023; White et al. 2023) to promote continuous and effective information exchange between teachers and GAI; guiding GAI to propose open-ended questions to foster the development of students’ critical thinking.

Fourth, integrate various teaching methods and tools to build an intelligent learning environment. In addition to applying conversational GAI tools, it is also essential to actively explore other specialized GAI tools (such as customized GAI tools for writing and speaking training) and combine them with other information technology teaching tools, comprehensively utilizing various technological means to construct an intelligent learning environment (Chen et al. 2020). By carefully designing prompts, facilitate the integration of GAI with learning management systems and online resources, providing integrated learning support for students, and improving their learning outcomes and efficiency.

Fifth, pay attention to potential risks, adhere to ethical norms, and cultivate awareness and ability to use GAI correctly. Although GAI’s text and audio-video processing capabilities seem powerful, attention must also be paid to the potential risks associated with these applications (Akinwalere & Ivanov 2022; Lu Yu et al. 2023). In GAI applications, teachers should educate students to adhere to academic ethics and moral norms, prevent misuse of technology, and also emphasize cultivating students’ awareness of correctly using GAI tools, highlighting the importance of copyright, privacy, and other issues, and being aware of the potential negative emotional impacts on students as classroom teaching shifts from human interaction to human + machine interaction (Jiao Jianli, Chen Ting 2023).

With the continuous advancement of information technology, the application of GAI in higher education is gradually changing traditional teaching models and processes. The impact of GAI on university English teaching and learning may be profound and systematic, but how it will specifically influence remains to be further observed and studied.

This article was published in the “Frontiers of Foreign Language Education Research” in 2024, issue 4, pages 41-50. Due to space limitations, footnotes and references have been omitted.

Wang Haixiao, 2024, Exploring the Application of Generative AI in University English Teaching Reform—A Case Study of the Teaching Reform Practice of the “General Academic English Writing” Course [J], “Frontiers of Foreign Language Education Research” (4): 41-50.