Guiding Generative AI with GSE Learning Objectives

Exploring Principled, Effective, and Accessible ELT Material Design

The Core Challenge

Exploring the bridge between AI potential and pedagogical principle.

🤔 The Problem

Generative AI tools offer immense possibilities for English teaching, but many implementations can lack pedagogical grounding. This may result in generic, misaligned, or ineffective learning content that doesn't meet specific student needs.

4K+

GSE Descriptors

0

AI Tools Tested

This application explores aligning ChatGPT, Copilot & Claude with the Global Scale of English (GSE) framework for a systematic approach to ELT material design.

Application Method

A practical approach to testing and applying AI tools.

  • Exploratory Testing

    Iteratively refining prompts based on specific GSE learning objectives to test the AI's ability to generate level-appropriate content.

  • Comparative Analysis

    Evaluating outputs from different GenAI tools for consistency, quality, accuracy, and adherence to the prompt's pedagogical constraints.

  • Contextual Review

    Connecting the practical application to current ideas in EdTech and language teaching to ensure it's relevant and informed.

ChatGPT
(OpenAI)
Copilot
(Microsoft)
Claude
(Anthropic)

Key Observations

How explicit pedagogical guidance can lead to better results.

Using explicit GSE descriptors in prompts appears to improve the quality and pedagogical alignment of GenAI responses.

GSE-aligned prompts appeared to yield:

  • More Level-Appropriate Content: Vocabulary & grammar seemed better matched to the target proficiency.
  • Better Task Matching: Activities felt more relevant to the specific skill.
  • Increased Consistency: Outputs were generally more reliable across different AI tools.

Example GSE Descriptor:

Skill: Speaking
CEFR: A2 / GSE: 30

"Can give a short, basic description of their home, family and job, given some help with vocabulary."

Practical Application

Example prompts for direct use and adaptation.

Example 1: Reading Comprehension (A2) 

Act as an English language teaching material designer. Create a short reading text (around 100 words) suitable for an A2 level learner.

The text should be based on the GSE learning objective (Reading, 34): "Can understand clearly written, straightforward instructions on how to use a piece of equipment."

After the text, create 3 simple multiple-choice questions to check comprehension.

Example 2: Writing Task (B1) 

As an expert in ELT, design a writing task for a B1 level student.

The task must align with the GSE learning objective (Writing, 46): "Can write simple informal emails/letters and online postings giving news or opinions."

The task is to write a short email to a friend (about 50-80 words).

Provide a clear scenario for the email (e.g., inviting them to a weekend trip) and include 3 bullet points of information they must include in their email.

Webinar in Action

Watch a practical demonstration of these principles at work.

In this recorded session, I walk through the process of using GSE-aligned prompts to generate high-quality, level-appropriate ELT materials with various Generative AI tools.

Conclusion & Impact

Exploring how educators can leverage AI responsibly and effectively.

Key Takeaways

  • Powerful Enhancement: GenAI tools, when guided, can be powerful aids for ELT material creation.
  • Pedagogical Soundness: Framework integration helps ensure outputs are coherent and educationally valuable.
  • Teacher Agency is Key: The educator's role in prompt design and evaluation remains critical for responsible AI integration.

Future Outlook

The future may lie in developing more intuitive interfaces and AI training that inherently understands pedagogical frameworks. This could lower the barrier to entry for educators and enable even more dynamic, personalized learning experiences.

This systematic approach can help empower educators to move from being simple users of AI to becoming architects of AI-driven learning.