TLDR: This research explores “vibe coding” – using natural language to create software with AI – in English as a Foreign Language (EFL) education. It introduces a “human-AI metalanguaging framework” with three dimensions: talking to AI (prompt engineering), talking through AI (negotiating authorship), and talking about AI (mental models of AI). Through a case study of two students, the study reveals that explicit instruction in these three dimensions is crucial for successful AI-powered application development and enhanced language learning, highlighting the impact of prompt structure and AI perception on project outcomes.
In an increasingly AI-powered world, a new approach called “vibe coding” is emerging, allowing individuals without traditional coding knowledge to create software applications using natural language. This innovative method holds significant promise for English as a Foreign Language (EFL) education, empowering students to design tools that address real-world language learning challenges.
A recent research paper, A Vibe Coding Learning Design To Enhance EFL Students’ Talking To, Through, and About AI, delves into the piloting of vibe coding within EFL contexts. Authored by David James Woo, Kai Guo, and Yangyang Yu, the study introduces a comprehensive human-AI metalanguaging framework designed to understand how EFL students interact with AI during this creative process.
Understanding the Human-AI Metalanguaging Framework
The framework proposes three interconnected dimensions through which students engage with AI:
- Talking to AI (Prompt Engineering): This refers to the art of crafting effective instructions or prompts for an AI. It demands metalinguistic and metacognitive skills as students learn to communicate precisely with an unfamiliar digital entity.
- Talking Through AI (Negotiating Authorship): This dimension explores the complex process of determining who holds authorship in a human-AI co-creation. It involves understanding the balance between human input and AI-generated content.
- Talking About AI (Mental Models of AI): This focuses on students’ internal perceptions and understandings of what AI is and how it functions. These mental models influence how they interact with and perceive the AI’s role.
The researchers conceptualize AI as a “beneficial languaging machine” that not only stimulates students’ language production but also encourages reflection and self-regulation, thereby enhancing their AI literacies, agency, creativity, and engagement.
The Pilot Study: Two Students, Two Approaches
To explore these interactions, the study conducted a four-hour workshop with two Form Three (Grade 9) students, Student A and Student G, from a Hong Kong secondary school. Both students were tasked with designing an application to address authentic EFL writing challenges, using backward design principles.
Student A’s Systematic Success
Student A approached the task with meticulous planning. Her design goal was to help Form 4 students practice process writing independently. She developed extensive, structured prompts, totaling over 2,200 words across four prompts, often preparing them offline before pasting them into the AI application creator. Her strategy involved iteratively expanding a single prompt by adding sections and information. This systematic approach, viewing AI as a precise tool requiring detailed instructions, led to the successful creation of a functional app called “Your Friendly Learning Buddy.” This app featured writing input boxes, word count display, and generated content, language, and organization scores with qualitative feedback, closely aligning with her initial design.
Student G’s Creative Challenges
In contrast, Student G envisioned a highly engaging, gamified app called “Brainrot while learning!” Her prompts were conversational and brief, totaling only 166 words over eight prompts. She viewed AI as a collaborative, creative partner, often expressing emotional responses to its outputs. However, Student G encountered significant technical difficulties, including platform access issues and message limits, which severely constrained her progress. Her submitted app, “Writing Quest,” ended up as a static webpage with a three-section layout, a significant departure from her ambitious, immersive video game concept. This case highlighted a misalignment between her creative design, her conversational prompting style, and the technical realities, resulting in a less functional outcome.
Also Read:
- Unpacking Writing Scores: A New Approach to Explaining Automated Evaluations
- AI Course Assistants: Bridging Support Gaps and Uncovering Interaction Dynamics in Computer Science Education
Key Insights for Educators
The contrasting experiences of Student A and Student G offer valuable lessons for educators:
- The way students “talk to, through, and about AI” directly impacts their vibe coding outcomes.
- Explicit scaffolding in metalanguaging is essential. Teachers should guide students in structured prompt engineering, facilitating critical discussions about authorship, and developing a rich vocabulary to articulate their mental models of AI.
- Student A’s success underscores the importance of planning and precision in prompt engineering, suggesting pedagogical activities that focus on structuring complex requests and providing one-shot examples.
- Student G’s experience highlights the need to balance ambitious design thinking with pragmatic technical planning, perhaps by starting with simpler projects to build foundational vibe coding skills.
Ultimately, this research demonstrates vibe coding’s potential to transform EFL education, provided that instruction explicitly addresses the intricate interplay of how students communicate with, negotiate with, and understand artificial intelligence.
