Designing Conflict-Based Communicative Tasks in TCFL with ChatGPT: A Process Analysis

1. Introduction

The advent of Artificial Intelligence (AI) is transforming various fields, including language teaching and learning. Applications like machine translation (e.g., DeepL), grammatical error correction (e.g., Grammarly), and text-to-speech synthesis (e.g., TTSmaker) are now commonplace. The late 2022 launch of ChatGPT, an AI-powered virtual assistant by OpenAI, has captured significant attention due to its remarkable information processing capabilities. This context necessitates a closer examination of AI applications in language didactics, specifically their impact on both teaching and learning processes.

This article focuses on analyzing the process of designing conflict-based communicative tasks for a university-level Oral Expression course in Teaching Chinese as a Foreign Language (TCFL) with the assistance of ChatGPT. It adopts a descriptive research perspective to expose the salient characteristics of the teacher-AI interaction and to illuminate its impacts on finalizing the teaching program design.

2. Context and Research Framework

2.1 Context of the Study

The study is situated within the development of a teaching program for a TCFL oral expression course at the university level. The core pedagogical strategy involves designing communicative tasks rooted in conflict scenarios to stimulate interactive dynamics among learners and foster the development of oral interactional competence.

2.2 Research Questions and Methodology

The research is guided by two primary questions:

1. How is the use of ChatGPT manifested during the process of elaborating conflict-based communicative tasks?
2. To what extent does its use influence the final teaching program?

The methodology is qualitative and descriptive, analyzing the corpus of interactions between the teacher-researcher and ChatGPT during the task design phase. The analysis aims to identify patterns, strategies, and decision-points within this human-AI collaborative design process.

3. Theoretical Framework

3.1 Communicative Tasks and Conflict Theory

A communicative task is defined as an activity where meaning is primary, there is a communicative goal to achieve, and success is evaluated in terms of outcome. Integrating conflict theory into task design introduces an element of cognitive and social dissonance—disagreements, differing perspectives, or problem-solving scenarios—that compels learners to negotiate meaning, justify opinions, and employ persuasive language, thereby deepening engagement and linguistic output.

3.2 Criteria for Task Elaboration

Key criteria considered during task design include:

• Authenticity: The conflict scenario should reflect real-world situations learners might encounter.
• Linguistic Appropriateness: The task must be calibrated to the learners' proficiency level (e.g., B1/B2 CEFR).
• Interactional Goal: Clear objective for the interaction (e.g., reach a compromise, persuade, solve a dilemma).
• Cognitive Demand: The conflict should require analysis, evaluation, and creative thinking.

4. Analysis of Teacher-ChatGPT Interaction

4.1 Manifestation of ChatGPT Use

The teacher used ChatGPT as a collaborative design partner. Prompts were structured to:

1. Generate Ideas: "Suggest 5 conflict scenarios for intermediate Chinese learners about sharing an apartment."
2. Refine Language: "Rephrase this task instruction to make it clearer for students."
3. Develop Content: "Provide a sample dialogue for this 'cultural misunderstanding at a dinner' scenario."
4. Evaluate & Critique: "Review this task outline and identify potential pitfalls for student engagement."

The interaction was iterative, with the teacher guiding, filtering, and adapting ChatGPT's outputs.

4.2 Impact on Final Teaching Program

ChatGPT's influence was observed in:

• Increased Ideational Variety: The AI proposed a broader range of conflict scenarios than initially conceived by the teacher alone.
• Accelerated Prototyping: Rapid generation of task drafts and materials reduced initial design time.
• Linguistic Scaffolding: Provided ready-made lexical chunks and sentence patterns relevant to conflict negotiation.
• Potential for Generic Outputs: Some suggestions lacked cultural specificity or pedagogical nuance, requiring significant teacher intervention to contextualize.

5. Technical Details and Analytical Framework

The analytical framework for evaluating the AI-assisted design process can be conceptualized through an input-process-output model with a feedback loop.

Process Evaluation Metric: A simple scoring mechanism can be used to assess the utility of each AI interaction. Let $U_i$ represent the utility of the i-th ChatGPT output, scored by the teacher on a scale from -1 (counterproductive) to +1 (highly useful). The average utility $\bar{U}$ for a design session is:

$$\bar{U} = \frac{1}{n}\sum_{i=1}^{n} U_i$$

Where $n$ is the number of significant AI interactions. A positive $\bar{U}$ indicates net positive assistance.

Interaction Pattern Classification: Interactions were coded as:

1. Divergent Ideation (DI): AI expands possibilities.
2. Convergent Refinement (CR): AI helps specify and improve.
3. Linguistic Generation (LG): AI produces language samples.
4. Pedagogical Critique (PC): AI evaluates task structure (limited).

6. Results and Discussion

Chart Description (Hypothetical): A bar chart titled "Frequency of ChatGPT Interaction Types During Task Design" shows the distribution. Divergent Ideation (DI) and Linguistic Generation (LG) are the most frequent interaction types, indicating ChatGPT's primary role as a brainstormer and language resource. Pedagogical Critique (PC) is the least frequent, highlighting the AI's current limitation in deep pedagogical analysis.

The analysis revealed that ChatGPT served most effectively as a catalyst and resource library, but not as a pedagogical expert. The teacher's role remained central in ensuring cultural authenticity, aligning tasks with learning objectives, and applying principles of second language acquisition (SLA). The final program was richer in scenario variety but required careful curation to maintain pedagogical coherence.

7. Case Study: Framework Application

Scenario: Designing a task for intermediate learners on "Negotiating Work Responsibilities."

1. Teacher Prompt (DI): "Generate 3 conflict scenarios between two colleagues in a Chinese office setting."
2. ChatGPT Output: Provides scenarios about unequal workload, missed deadlines, and credit for ideas.
3. Teacher Action (CR): Selects the "unequal workload" scenario and prompts: "List 5 key Mandarin phrases for politely complaining about workload and 5 for refusing a task."
4. ChatGPT Output (LG): Provides phrases like "我最近工作量有点大…" and "我可能暂时接不了这个任务…"
5. Teacher Synthesis: Integrates the scenario and phrases into a role-play task card, adding clear instructions and success criteria based on pedagogical goals.

This case illustrates the iterative, guided use of ChatGPT, where the AI supplies content that the teacher pedagogically frames.

8. Future Applications and Directions

• Specialized AI Tutors: Development of LLMs fine-tuned on SLA research and specific language corpora (e.g., TCFL learner interlanguage) to provide more pedagogically sound suggestions.
• Dynamic Task Generation: AI systems that can generate personalized conflict tasks based on a learner's profile, interests, and observed linguistic gaps.
• Multimodal Integration: Combining ChatGPT with image/video generation AI (e.g., DALL-E, Sora) to create immersive, scenario-based conflict prompts.
• Real-Time Interaction Analysis: Using AI to analyze recordings of learners performing conflict tasks, providing feedback on interactional strategies, turn-taking, and pragmatic competence.
• Ethical & Cultural Frameworks: Building guardrails and cultural modules into AI assistants to ensure generated conflicts are appropriate, constructive, and culturally informed.

9. References

1. Ellis, R. (2003). Task-based language learning and teaching. Oxford University Press.
2. Long, M. H. (2015). Second language acquisition and task-based language teaching. Wiley-Blackwell.
3. OpenAI. (2022). ChatGPT: Optimizing Language Models for Dialogue. https://openai.com/blog/chatgpt
4. Pica, T., Kanagy, R., & Falodun, J. (1993). Choosing and using communicative tasks. In G. Crookes & S. M. Gass (Eds.), Tasks and language learning: Integrating theory and practice (pp. 9-34). Multilingual Matters.
5. Warschauer, M., & Healey, D. (1998). Computers and language learning: An overview. Language Teaching, 31(2), 57-71.
6. Zhao, Y. (2023). The AI-Powered Language Teacher: A Framework for Integration. CALICO Journal, 40(1), 1-25.

10. Original Analysis & Expert Commentary

Core Insight: This study isn't about AI replacing teachers; it's about AI augmenting the creative and logistical dimensions of pedagogical design. The real story here is the emergence of the "teacher-as-curator-and-prompt-engineer." The value isn't in ChatGPT's raw output, which as the paper notes can be generic, but in a skilled educator's ability to frame prompts that extract pedagogically useful raw material and then refine it. This mirrors findings in creative industries using AI, where the human's role shifts from sole creator to strategic director (Ammanath, 2022).

Logical Flow & Strengths: The paper correctly identifies the AI's sweet spot: divergent ideation and linguistic scaffolding. By offloading the cognitive load of generating numerous scenario ideas and associated vocabulary, the teacher can focus on higher-order pedagogical tasks—structuring the interaction, setting appropriate goals, and integrating the task into a broader curriculum. This aligns with the concept of "distributed cognition," where tools handle routine cognitive tasks, freeing human intelligence for complex problem-solving (Hutchins, 1995). The descriptive methodology is appropriate for this nascent field, providing a rich, qualitative map of the interaction terrain.

Flaws & Critical Gaps: The analysis, while valuable, skims the surface of the prompt engineering process. What specific prompt structures yielded the best results? This is the new core competency for educators, akin to a programmer's skill. The paper also lacks a comparative analysis. How did the AI-assisted design process differ in efficiency, creativity, and outcome from a traditional, teacher-only or teacher-peer collaborative process? Furthermore, the ultimate impact—student learning outcomes—is absent. Do conflict tasks designed with AI lead to better oral interaction skills than those designed without? This is the crucial, unanswered question. The study, like many in EdTech, focuses on the tool's use by the teacher, not its final effect on the learner, a common pitfall noted by researchers like Selwyn (2016).

Actionable Insights: For language departments and educators: 1) Invest in prompt literacy training. Professional development should move beyond basic AI usage to advanced techniques for eliciting pedagogically robust content. 2) Develop shared prompt libraries. Create a repository of vetted, effective prompts for TCFL task design (e.g., "Generate a B1-level role-play conflict about [topic] incorporating phrases for [function]"). 3) Adopt a critical, iterative workflow. Use AI for the first draft, but mandate multiple rounds of human review focused on cultural nuance, pedagogical alignment, and avoiding AI bias or "smooth" but inauthentic language. 4) Initiate longitudinal studies. The field must move from process descriptions to outcome-based research. Partner with learning scientists to measure the efficacy of AI-co-designed materials on actual language acquisition metrics. The future belongs not to teachers who fear AI, but to those who learn to harness it as a powerful, if imperfect, co-pilot in the complex journey of language pedagogy.