Teachers iteratively refine LLM-simulated student ideas with classroom context before instructional use¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The article recommends that teachers treat LLM output as a starting point for co-development rather than finished material: through iterative prompting, "teachers can correct the output and clarify classroom contexts, including prior instruction, learning objectives, discussion norms, and students' existing knowledge." Teachers may refine learning objectives and re-prompt the LLM, generating new facilitation ideas. This positions teachers as collaborators and innovators who participate in creating, evaluating, and refining GenAI-enabled learning activities, and it requires teacher oversight to align simulated ideas with the lesson's knowledge scope.

Design Implications¶

Context¶

Requirements¶

• Teacher oversight and idea co-development to align LLM-generated ideas with the targeted lessons' knowledge scope
• Professional development introducing teachers to responsive teaching vision and guidance on critiquing AI simulations

Constraints¶

• Some LLM-generated prompts were overly leading, redirecting conversations to AI-chosen topics
• LLM ideas may use advanced or unrealistic vocabulary unnatural for students

Target Learners¶

• in-service science teachers
• pre-service teacher education novices

Target Learning Goals¶

• anticipating student ideas
• planning responsive instructional moves

Affordances¶

• Responsive Teaching Framework Nguyen Cao

Related Strategies¶

• Use LLM-generated lessons within a human-AI collaborative workflow where educators review and refine AI drafts
• Treat AI-generated lesson plans as drafts requiring teacher evaluation, adaptation, and pedagogical judgment
• Integrate LLM-based CRP support tools within professional development to encourage teacher reflection beyond template-based outputs
• Use an AI-assisted vibe coding workflow with human pedagogical control to move educational tools from idea to classroom within days
• Co-create micro-intervention repertoires with generative AI through a seed-request-refine prompt sequence, with the instructor curating output
• Use generative AI as a brainstorming partner for divergent ideation while treating didactic transformation as a human responsibility
• Educators gamify learning materials or create learning games using generative AI with iterative, phased prompting

Examples¶

-

Key Sources¶

• Nguyen, H., & Cao, J. (2026). Exploring the Capacity of Large Language Models to Simulate Students' Scientific Thinking: Insights for Responsive Teaching. Journal of Science Education and Technology. https://doi.org/10.1007/s10956-026-10333-5