Using Large Language Models to Analyze Technology Integration in K–12 Canadian Curricula

• In Event: AERA Roundtable Session Wednesday 1:45 pm - Gold 3
  In Roundtable Session: Integrating Artificial Intelligence in Teaching and Learning

Abstract

This study utilizes large language models (LLMs) to investigate the discussion of technology in approximately 1,400 K–12 Canadian curriculum documents over two decades. We analyze technology-related content at provincial, subject, and grade levels over time. We use ChatGPT-4o to classify each page of a document based on whether it references technology, followed by AI-assisted thematic analysis to uncover key themes. Validation with human experts demonstrates high inter-rater reliability (MKappa = 0.90). Preliminary findings provide a preview of insights to be gained from the complete database. This project informs educators and policymakers about how technology is integrated across educational contexts, highlighting gaps in instruction, access and usage. It also demonstrates a novel, replicable method for large-scale qualitative analysis using LLMs.

Authors

• Esther Yu, University of Toronto - OISE

• Calpanaa Jegatheeswaran, University of Toronto

• Sumayya Saleem, University of Toronto - OISE

• Elizabeth Dhuey, University of Toronto - Scarborough

• Michal Perlman