Pre-service science teachers' adaptive pedagogical reasoning in AI-supported lesson plans: a case with chemical equilibrium

Although generative artificial intelligence (AI) is increasingly being integrated into lesson planning in teacher education, little is known about how pre-service science teachers adapt AI-generated content through pedagogical reasoning for contextually responsive science teaching. This study explored how AI influences pre-service science teachers' (PSSTs) adaptive pedagogical reasoning (APR) when planning lessons on chemical equilibrium in resource-constrained contexts. Drawing on the Refined Consensus Model of Pedagogical Content Knowledge, APR was conceptualised across four dimensions: identification and anticipation of students' thinking, adaptivity in lesson design, justification of instructional decisions, and alignment with learning goals and differentiation. An exploratory mixed-methods case study was conducted with 19 third-year PSSTs in a South African university. Participants developed traditional and AI-supported lesson plans using the Rationale for Lesson Design framework. Data were analysed using an APR rubric and Rasch analysis. Findings showed that AI-supported planning enhanced anticipation of students' thinking, justification of instructional decisions, and alignment with learning goals and differentiation. However, adaptivity in lesson design remained limited. The study suggests that while AI can scaffold deeper pedagogical reasoning, teacher education programmes should explicitly support adaptive instructional decision-making.