Research-backed AI prompts for educators

Give the AI a role and context

Tell the AI who it should be (e.g., "You are an experienced Year 10 biology teacher") and provide your subject, year level, and class size. All three major AI providers (OpenAI, Anthropic, and Google) list role assignment and contextual detail as foundational prompting strategies. Mollick and Mollick (2023) show that education-specific context produces significantly more usable classroom outputs. As Andrej Karpathy, co-founder of OpenAI, recently put it: the real skill is not "prompt engineering" but "context engineering", the art and science of filling the context window with the right information so the model can do its best work.

Specify the format you want

State whether you want bullet points, a table, a numbered list, or prose, and include a target length. OpenAI, Anthropic, and Google all identify output format specification as a core prompting technique. In education, this matters because outputs need to fit specific classroom activities, handouts, or assessment templates.

Add constraints

Include limits like word count, reading level, or curriculum alignment (e.g., "aligned to the Australian Curriculum"). Google's Gemini prompting guide recommends using constraints to narrow scope, and Mollick and Mollick (2023) demonstrate that pedagogical constraints produce outputs better suited for direct classroom use.

Iterate and refine

Treat the first response as a draft. Follow up with refinements ("make it shorter", "add a Bloom's Taxonomy tag to each question") rather than starting over. In Co-Intelligence (2024), Mollick describes effective AI use as an ongoing conversation. Both Anthropic and OpenAI recommend breaking complex tasks into chained steps for better results.

Always review before using with students

AI outputs can contain factual errors, cultural assumptions, or inappropriate content. Always review and adapt before sharing with students or colleagues. UNESCO's Guidance for Generative AI in Education and Research (2023) and Harvard's AI Pedagogy Project both emphasise human oversight of all AI-generated educational material.