Co-constructing adaptive lesson plans with GenAI: Pre-service teachers' Intelligent-TPACK and prompt engineering strategies

Abstract

Generative Artificial Intelligence (GenAI) technologies present new opportunities for teachers to design adaptive and student-centered instruction. However, the educational value of GenAI depends not only on technical usage but also on teachers' ability to formulate pedagogically meaningful prompts. Prompting strategies are not isolated from teachers′ prior knowledge and skills. Less is known about how pre-service teachers′ AI-related knowledge influences prompt engineering strategies, in turn leading to meaningful adaptive lesson plans. Considering this gap, we design an instructional task for pre-service teachers to generate adaptive lesson plans with the help of GenAI. Prior to the task, we collected data about their AI-related skills, namely AI literacy and Intelligent-TPACK. The prompts were qualitatively analyzed based on the phases of the Knowledge Construction (KC) Framework. Then, we explored the pedagogical value of adaptive lesson plans through a rubric in terms of three indicators: student agency, adaptive strategies, and flexible tools. PLS-SEM analysis revealed that as long as pre-service teachers have AI-specific technological and pedagogical knowledge, they formulate higher phases of prompts based on the KC framework. Our analysis showed that prompts from higher phases generated more adaptive lesson plans in terms of student agency, adaptive strategies, and flexible tools. We also found an indirect effect of Intelligent-TPK on adaptive lesson plans. This study highlights that effective prompt engineering is a pedagogical act shaped by teachers’ knowledge, not merely a technical command. It also underscores the importance of embedding AI-specific pedagogical training in teacher education. By conceptualizing prompts as epistemic moves, we offer new insights into how teachers and GenAI can collaborate to produce responsive and inclusive learning experiences.