Short answer

Leverage AI, specifically LLMs, to assist in the creation of adaptive and effective educational content, potentially surpassing human-generated materials in specific contexts.

Field
Innovation & Design
Source
British Journal of Educational Technology (2025)
Method
Cognitive Task Analysis and LLM-based content generation evaluation
Evidence
Strong effect

Large Language Models (LLMs), when provided with original curriculum materials and expert-informed prompts, can create supplementary educational content (scaffolds) that are rated higher than those developed by experienced teachers. This innovation & design research insight is drawn from a 2025 study published in British Journal of Educational Technology. Using Cognitive task analysis and llm-based content generation evaluation, researchers explored how this design variable affects real-world outcomes. The key design takeaway: Leverage AI, specifically LLMs, to assist in the creation of adaptive and effective educational content, potentially surpassing human-generated materials in specific contexts.

Study
Innovation & DesignNew This WeekStrong effect

LLMs can generate superior math curriculum scaffolds compared to expert teachers

Large Language Models (LLMs), when provided with original curriculum materials and expert-informed prompts, can create supplementary educational content (scaffolds) that are rated higher than those developed by experienced teachers.

British Journal of Educational Technology · 2025

01

Key Findings

  • 01A three-stage process for curriculum scaffolding (observation, strategy formulation, implementation) was identified from expert teachers.
  • 02LLMs can generate curriculum scaffolds that are rated significantly higher than those created by expert teachers.
  • 03Providing LLMs with the original curriculum materials and an expert-informed prompt yielded the best results.
02

Application

Design takeaway

Leverage AI, specifically LLMs, to assist in the creation of adaptive and effective educational content, potentially surpassing human-generated materials in specific contexts.

How to apply

Develop and test LLM-powered tools that can generate differentiated learning materials or practice exercises based on core curriculum content and specific student needs.

Project actions

  • 01When designing educational tools, consider how AI can personalize content.
  • 02Focus on how to effectively prompt AI to achieve desired educational outcomes.
03

Method & Evidence

AimCan large language models (LLMs) effectively support middle school math teachers in creating high-quality curriculum scaffolds that enhance student access and engagement?
MethodCognitive Task Analysis and LLM-based content generation evaluation
ProcedureExpert teachers were interviewed to identify their curriculum scaffolding processes. This information was used to develop and test three different LLM approaches for generating math warm-up tasks. The LLM-generated tasks were then evaluated by teachers against criteria such as alignment to learning objectives, accessibility for lower-performing students, and overall teacher preference.
ContextMiddle school mathematics education

Variables

IVLLM approach (e.g., prompt quality, context provided)
DVQuality of generated curriculum scaffolds (rated by teachers on alignment, accessibility, preference)
CVSubject matter (middle school mathematics), type of scaffold (warm-up tasks), teacher expertise
04

Strengths & Limitations

Strengths

  • +Direct comparison between LLM-generated and teacher-generated content.
  • +Incorporation of expert teacher insights into LLM prompt design.

Limitations

The effectiveness of LLM-generated content can depend heavily on the quality of the prompt and the specific LLM used. Teacher acceptance and training are also crucial factors.

Reliability & validity

The study's validity is supported by expert teacher ratings. Reliability could be enhanced by using a larger, more diverse group of raters and by standardizing the LLM prompting process across multiple trials.

Think critically

To what extent should AI replace human expertise in curriculum development, and what are the potential long-term consequences for pedagogical innovation?

05

Design Principles

"AI-augmented content creation can enhance the quality and accessibility of educational resources."

This research highlights a significant advancement in educational technology, suggesting that AI can augment the capabilities of educators. By automating the creation of tailored learning materials, LLMs could free up teacher time and improve the accessibility and effectiveness of curricula for a wider range of students.

06

What This Means for Your Design

Computers can now help teachers make better math lessons by creating extra activities that are easier for all students to understand.

How to use in your project

  • 1.Use this research to justify the use of AI tools in your design process for educational products.
  • 2.Cite this study when discussing the potential of AI to create adaptive learning content.
07

Add to My Project

08

Quick Cite

Paragraph starter

This research demonstrates that Large Language Models (LLMs) can be powerful tools for generating educational content. Specifically, when provided with original curriculum materials and expert-informed prompts, LLMs produced math warm-up tasks that were rated higher by teachers than those created by human experts in terms of alignment to learning objectives and student accessibility. This suggests a significant opportunity for AI to support educators in creating effective and inclusive learning experiences.

09

Source

British Journal of Educational Technology

Scaffolding middle school mathematics curricula with large language models

journal · 2025

View source

Questions About This Research

What does the research say about llms can generate superior math curriculum scaffolds compared to expert teachers?
Leverage AI, specifically LLMs, to assist in the creation of adaptive and effective educational content, potentially surpassing human-generated materials in specific contexts. Evidence: British Journal of Educational Technology (2025).
Why does "LLMs can generate superior math curriculum scaffolds compared to expert teachers" matter for design?
This research highlights a significant advancement in educational technology, suggesting that AI can augment the capabilities of educators. By automating the creation of tailored learning materials, LLMs could free up teacher time and improve the accessibility and effectiveness of curricula for a wider range of students.
How can designers apply this research?
Leverage AI, specifically LLMs, to assist in the creation of adaptive and effective educational content, potentially surpassing human-generated materials in specific contexts.
What were the main findings?
A three-stage process for curriculum scaffolding (observation, strategy formulation, implementation) was identified from expert teachers.. LLMs can generate curriculum scaffolds that are rated significantly higher than those created by expert teachers.. Providing LLMs with the original curriculum materials and an expert-informed prompt yielded the best results.
What research method was used?
Cognitive Task Analysis and LLM-based content generation evaluation.
How strong is the evidence?
Evidence strength is rated Strong effect, based on a 2025 journal from British Journal of Educational Technology.
What should I do differently in my next project?
Develop and test LLM-powered tools that can generate differentiated learning materials or practice exercises based on core curriculum content and specific student needs.
What are the limitations?
The study focused on middle school math warm-up tasks; generalizability to other subjects, age groups, or types of educational content may vary. The 'expert-informed prompt' is a critical but potentially complex element to replicate.