1 Introduction

Output oriented approach (POA) is a Chinese foreign language education theory that places high demands on instructional design and poses challenges for teachers. ChatGPT and other major language models (LLMs) have shown great potential in foreign language teaching, but current research is mostly macro level exploration and lacks specific teaching design examples.

The Motivation of Human Computer Collaborative Design in POA Teaching

POA instructional design faces four major challenges: insufficient creativity in output scenario design, high requirements for preliminary analysis of output goals, high complexity in facilitating activity design, and tight time constraints in evaluation stage design. LLMs such as ChatGPT have the potential to empower POA instructional design, providing knowledge support, adaptive and differentiated teaching, and in-depth text analysis capabilities. However, there is currently a lack of research on integrating LLMs into real-life instructional design contexts. This study aims to explore the application of ChatGPT in various aspects of POA teaching design, leverage the advantages of human-computer collaboration, enhance the scientific and creative nature of teaching design, and empower college English teaching.

3 Case Studies of Human Computer Collaborative POA Teaching Design

This article introduces how to use ChatGPT-3.5 in collaboration with teachers to complete the Output Oriented Approach (POA) teaching design in the college English reading, writing, and translation courses for non English major first-year undergraduate students in "Double First Class" universities. The teaching material is the "Green Life" unit, which aims to enhance students' English reading and writing skills and environmental awareness through the overall output task of "Writing a Green Life Guide". ChatGPT provides various solutions in the design of production scenarios to help teachers save time in finding creative ideas, but teachers need to further refine the scenario design to ensure the authenticity of communication scenarios. In the stage of setting output goals, ChatGPT assists teachers in clarifying teaching priorities, bridging student gaps, analyzing textbooks from three dimensions: content, language, and structure, and proposing teaching priorities. In facilitating activity design, ChatGPT assists teachers in extracting target language items and generating reference answers or examples that can be used in the classroom to improve design efficiency, but teachers still need to build a facilitating activity chain. In the evaluation focus mining, ChatGPT analyzes student works, points out common problems, and teachers determine the evaluation focus and typical samples based on this, and conduct teacher-student cooperative evaluation. Throughout the process, ChatGPT provides assistance to teachers, but teachers still need to analyze, judge, integrate, and optimize information based on the foundation provided by ChatGPT.

COPILOTS human-machine collaboration mode in POA teaching design

This article introduces the COPILOTS human-machine collaboration mode in POA instructional design, emphasizing the role of LLMs as "co pilots" in instructional design and the action chain of alternating assignments with teachers to complete the entire instructional design process. The COPILOTS model involves four stages: creating output scenarios, setting output goals, facilitating activity design, and mining evaluation focuses, which are achieved through human-machine collaborative design routes. LLMs are responsible for initially creating output scenarios, optimizing them by teachers, generating output samples and evaluating trial outputs, and integrating information to confirm output goals by teachers. In the stage of facilitating activity design, LLMs assist individual units in facilitating activity design and teachers organize and build it. In the evaluation process, LLMs analyze and explore common issues in student output samples, and teachers select focus and typical samples. This model requires teachers to possess POA theory literacy, AI technology literacy, and higher-order thinking literacy to ensure the smooth implementation of human-machine collaboration.

5 Conclusion

This chapter summarizes the research on applying ChatGPT to the design of college English POA teaching, emphasizing its advantages in enhancing the scientificity and accuracy of teaching design, broadening teachers' thinking, reducing workload and increasing efficiency, and lowering the threshold for POA teaching design and application. At the same time, the limitations of ChatGPT in instructional design were pointed out, including the randomness of responses and the issue of rough processing of information. It emphasized the irreplaceability of teachers in POA instructional design and the need for comprehensive analysis and deep processing of LLMs' content to give instructional design vitality.