AI Prompts in Plurilingual, AoA Teaching

subheader extendedexp | Action-oriented Approach|Plurilingualism|Tech-mediated

Recap 📝

Before you start exploring this infosheet, you might want to check:

  1. Introduction to Plurilingualism
  2. Introduction to the Action-oriented Approach
  3. Understanding AI Prompts
  4. AI Prompts for Plurilingual Pedagogy

A cautionary note: Large Language Models (LLMs) learn to generate text from extensive self-training and specific data, but their output requires continual critical evaluation due to potential biases or inaccuracies in their training. Well-crafted AI prompts, using critical AI digital literacy, guide these models to create language teaching resources, offering efficiency and new ideas for educators.

Prompt engineering for plurilingual, action-oriented learning 💡

Generative AI tools can support lesson planning, but only if prompts are structured effectively. Teachers must be critical of the output received, learn to prompt iteratively & craft prompts that:

  • Are specific, fully detailed, critical, and ongoing (keep prompting until you get what you need)
  • Integrate plurilingualism (e.g., encouraging translanguaging, comparing languages and cultures, leveraging students’ linguistic and cultural repertoires).
  • Align with action-oriented approaches (e.g., scenario-based, real-world communication).
  • Correspond to students’ needs, languages, cultures, and specific contexts.

Watch: Using Generative AI and prompt engineering strategies for lesson planning

Following the example in the video, you can use the following prompt template to create a task or scenario adapted to your specific context:

Act as [role, e.g. an experienced language educator], create a [duration] [task or scenario] for [level] [target language, e.g. Spanish, French, German]. The [task or scenario] must be designed under an [pedagogical approach, e.g. action-orientation] and incorporate [plurilingual strategies, e.g., translanguaging, comparing languages]. Learners should use the [target language, e.g. Spanish, French, German] language resources at a CEFR level [e.g. A1, A2, B1, B2, C1, C2] to [outcome]. Include [output requirements, e.g., can-do statements, CEFR descriptors] in your output.

Your turn 🫵

  • Use a generative AI tool such as ChatGPT, Copilot, or Gemini and one of the templates above to create a task or scenario of your choice according to your teaching context. 
  • Think of the task or scenario generated by the generative AI tool. Does it align with your expectations while adhering to plurilingual and action-oriented principles?

Note: Prompting engineering is a craft; you will probably not obtain your desired results on the first try. You will need to try different prompts and use different prompting strategies. You can also use frameworks such as C-TOE (Context, Task, Output, Example) and CIDI (ContextInstructionsDetailsInput) to structure your prompts.

Reflection questions 💭❓

  • What challenges do you anticipate from using a generative AI tool as a co-designer for your learning experiences?
  • Does the process of designing a task or scenario with generative AI increase in complexity or make it simpler? In what way(s)?
  • To what extent do you think the results provided by a generative AI can capture the nuances of tasks and scenarios designed with a plurilingual and action-orientation lens?

More useful resources ➕

AALE Toolkit Home | Action-oriented Approach | Plurilingualism | Tech-mediated | AALE Infosheets Exploration