Prompt Guide 101 for Teachers (K–12 & Higher Education)

Step‑By‑Step Prompt Writing Framework

Use this five‑step framework to craft prompts that consistently deliver useful results. Each step includes a template you can fill in.

Step 1: Define the goal

Decide what you want the AI to produce. Be specific—do you need a lesson outline, a quiz, examples for differentiation, or a parent email? Clear goals reduce the risk of vague outputs or hallucinations.

Template:

Example:

Goal: Draft a 20‑minute reading comprehension exercise based on the short story "The Gift of the Magi" for grade 8 students.

Step 2: Provide context

State who the output is for (grade level, subject) and any relevant background (learning objectives, student strengths or needs). Including context helps the model align with your classroom reality.

Template:

Example:

Step 3: Give clear, specific instructions

Detail the tasks or components you need. If multiple parts are required (e.g., warm‑up, lecture outline, assessment), list them. Use bullet points or numbered lists for clarity. If you have constraints (time limits, vocabulary level, format), spell them out.

Template:

Example:

Step 4: Set the output format

Indicate the structure you expect—bulleted list, table, rubric, paragraph, or script. Use cues like “Output as a table with columns X and Y” or “Provide the answer in a JSON object.” This primes the model.

Template:

Example:

Step 5: Test and revise

After running your prompt, evaluate the response for accuracy, appropriateness and tone. If the output is off-target, tweak your instructions, provide more context, or adjust the temperature. Consider lowering the temperature for factual tasks or raising it for creative brainstorming. Repeat until satisfied.

Template:

Prompt Examples for Classroom Use

Below are ready‑to‑use examples illustrating how to apply the framework across grade levels, subjects and tasks. Feel free to customize them.

Lesson plan creation

• Elementary (Grade 2 science – Plant growth)
  Prompt:

  You are an enthusiastic grade 2 science teacher. Create a 30‑minute lesson plan introducing plant growth. Include a short story to engage students, a simple hands‑on activity (planting seeds in cups) and three comprehension questions. Output as a bulleted list.

• Middle school (Grade 7 mathematics – Fractions)
  Prompt:

  Act as a math tutor for grade 7 learners. Design a 45‑minute lesson plan on adding and subtracting fractions with unlike denominators. Provide a warm‑up problem, a mini‑lecture outline, collaborative practice tasks and a brief quiz. Include differentiation ideas for advanced learners and those who need extra support.

• High school (Grade 10 history – World War II)
  Prompt:

  You are a high‑school history teacher preparing a 50‑minute lesson on the causes of World War II for grade 10 students. Create an outline with learning objectives, a five‑minute hook, three mini‑lectures (political, economic, social factors), and an exit ticket question. Ensure the reading level suits 15‑year‑olds.

• College (Introduction to psychology – Conditioning)
  Prompt:

  Serve as a college instructor for an introductory psychology course. Develop a 60‑minute lecture on classical vs. operant conditioning. Include definitions, key experiments, real‑world examples and a discussion question. Provide a slide outline and suggest a short video clip to support the lesson.

Differentiated instruction

You are a special education specialist supporting a mixed‑ability grade 4 classroom. Create three tiered practice activities to reinforce multiplication facts (×1–×12). One activity should be hands‑on for tactile learners, one digital game suggestion for tech‑savvy students, and one challenge problem set for advanced learners. Indicate the estimated time for each activity.

Rubric and assessment creation

Act as a high‑school English teacher. Design a rubric for grading a five‑paragraph argumentative essay on environmental conservation. Include criteria for thesis clarity, evidence/support, organization, grammar/style, and creativity. Use a 0–4 scale and define each level in one sentence.

Parent communication drafts

You are an assistant principal. Write a respectful email to parents informing them about an upcoming field trip to a science museum for grade 6 students. Include dates, cost, required permission slips, chaperone information and a paragraph on the educational value. Keep the tone warm and professional.

Creative writing support

As a creative writing coach, generate three story starters for a grade 8 creative writing class focusing on the theme of friendship. Each starter should be 1–2 sentences and leave room for students to imagine the rest. Make them diverse in tone (humorous, mysterious, heartfelt).

Student feedback or grading help

Act as a college biology professor providing formative feedback on student lab reports. Given a brief description of an experiment (you will paste it), produce constructive feedback highlighting strengths, areas for improvement, and a suggested grade out of 100. Use a supportive tone and avoid personal data.

STEM and humanities examples

• STEM – Coding task:

  Serve as a computer science tutor for first‑year university students. Explain the concept of recursion in Python, provide a simple code example, and craft three practice problems. Output should include syntax‑highlighted code blocks.

• Humanities – Literature analysis:

  You are a literature professor. Analyze the symbolism of the green light in F. Scott Fitzgerald’s “The Great Gatsby.” Provide three interpretations and cite textual evidence. Conclude with a discussion question for students.

Advanced Prompting Tips

Use roles/personas

Assigning a persona to the model helps set expectations and tone. Roles can range from “enthusiastic elementary teacher” to “skeptical librarian.” For example:

Act as a gentle math coach who explains concepts step‑by‑step and encourages mistakes. Teach long division to a group of 9‑year‑olds.

Personas can also mirror professional roles like “curriculum designer” or “special education coordinator.” Use them to shape voice, expertise, and style.

Prompt chaining (multi‑step tasks)

Prompt chaining involves breaking a complex task into smaller sub‑tasks and feeding the output of one step into the next. This method allows the model to build on previous reasoning, improving coherence and accuracy. For example:

1. Generate a list of five possible project themes for an English class.

2. For theme #2, outline the key concepts students should learn.

3. Create a 30‑minute lesson plan based on that outline.

Chaining works especially well for research projects, project‑based learning units and assessment design.

Using system prompts or memory

In the chat‑completion API, a system prompt sets overarching instructions for the entire conversation (e.g., “You are a friendly assistant who always cites sources”). When using platforms like ChatGPT with custom instructions or projects, you can upload documents and instructions that persist across sessions. According to Edutopia, ChatGPT’s Projects feature allows teachers to add reference materials and custom instructions so that the model remembers their preferred formats and styles. Use these features to establish classroom norms—such as always including learning objectives or aligning to standards.

Prompts for student use (safely)

When students interact with AI tools, set ground rules. Encourage them to:

• Ask the AI for brainstorming ideas, examples or explanations—not completed assignments.

• Use explicit constraints (grade level, word count) so outputs match their capacity.

• Evaluate AI responses critically and cross‑check facts.

• Avoid sharing personal or sensitive information (names, addresses).

• Use creative tasks (e.g., story starters, vocabulary practice) at higher temperatures to encourage exploration.

Prompt Do’s and Don’ts

Do’s

• Start with a clear goal and role: Determine what you need and who the AI should be (teacher, tutor, coach). This defines tone and output.

• Provide specific context: Grade level, subject, learning objectives and any constraints help tailor the response.

• Structure instructions: Use lists, numbered steps, and explicit constraints to reduce ambiguity.

• Set output formats: Prime the model with cues like “Here is a rubric…” to ensure structured results.

• Adjust temperature thoughtfully: Use low temperatures for factual, deterministic tasks and higher temperatures for brainstorming or creative writing.

• Evaluate and iterate: Treat prompt design as an iterative process. Test, review, refine and repeat.

• Cite or verify information: ChatGPT can hallucinate. Always verify facts and cite reputable sources when needed.

• Encourage critical thinking: Ask the model to show reasoning or explain step‑by‑step. Chain‑of‑thought prompting improves transparency.

Don’ts

• Don’t be vague: Broad prompts like “Write a report about history” yield general and potentially inaccurate responses. Specify topics, grade level and required sections.

• Don’t overload tokens: Remember the model’s context window. Avoid unnecessary whitespace and filler; be concise.

• Don’t rely on AI for sensitive decisions: Never use ChatGPT for grading high‑impact decisions (discipline, grading final scores) without human review. Models can reflect biases from their training data.

• Don’t treat AI outputs as fact: Generative AI can produce skewed or misleading content, particularly around topics of gender, race or politics. Use a “healthy hesitation” and apply your professional judgement.

• Don’t share personal data: Never include student names, addresses or medical information in prompts. Use generic or anonymized descriptions.

• Don’t set conflicting instructions: Avoid mixing multiple roles or contradictory constraints (e.g., “Be concise” and “Write in detail”). Clarify priorities.

Ethical use & bias considerations

Generative AI models can reflect and amplify societal biases. The MIT Sloan Teaching & Learning Technologies guide warns that AI tools have produced biased content related to gender, race and political affiliation. They also produce hallucinations—fabricated references or false information. To mitigate these risks:

1. Critically evaluate outputs: AI systems don’t “think.” They operate on patterns and probabilities, so you must review responses with human judgement.

2. Diversify your sources: Cross‑check AI‑generated information with expert‑vetted materials or peer‑reviewed publications.

3. Use retrieval‑augmented tools: Tools that retrieve information from trusted sources before generating answers can improve accuracy.

4. Employ structured prompts: Clear, specific instructions reduce the likelihood of hallucinations and unsupported claims.

5. Adjust temperature: Lower temperatures produce more consistent and factual outputs.

6. Engage students in discussions about bias: Use examples of biased outputs to spark conversations about media literacy and ethical AI.