Semantic Tagging of AI-Generated Science Texts to Identify Topics for Teaching STEM in K-12 Contexts

Abstract

As STEM education continues to play a crucial role in preparing students for future challenges, finding effective ways to improve its content and instructional delivery, especially in K-12 settings, requires ongoing attention and innovation. This study employed semantic tagging on AI-generated science texts to identify key semantic tags that can serve as topics for STEM instruction. Two datasets were analyzed using the English USAS semantic tagger and AntConc tool. The findings highlight topic-specific areas associated with animal and vegetable production (e.g., domestic animals, health and disease, farming and horticulture), increasing the students’ perspectives on agricultural practices, technological applications, system designs, and mathematical knowledge. The semantic tags identified can inform and enrich STEM curriculum design, offering educators a data-driven approach to selecting relevant and appropriate content for K-12 instruction. Overall, this study provides steps to examine other STEM-related texts for the development of curriculum design from a linguistic approach.