Mastering AI Prompt Engineering

What is AI Prompt Engineering?

Large Language Models (LLMs) like ChatGPT or Perplexity are trained on vast amounts of human conversation and writing. This allows them to predict the next word in a sentence, based on the patterns in the training data. LLMs rely on probability and context from your AI prompt to predict the best or most likely response.

For example, if you typed “I like to eat …”, it would rely on probability and context based on your prompt to predict the best or most likely response. In this example, chocolate or pizza would both be appropriate responses.

LLMs also learn which words are important in a sentence and which words are related. This is called Natural Language Processing, or NLP. NLP helps an LLM understand what you are asking and how to respond. 

Why AI Prompt Engineering Matters

LLMs use pattern recognition to respond to requests. Without human guidance, AI responses can be generic and lack context.

AI Prompts are your opportunity to introduce your educational expertise:

• Your student’s needs and performance
• Your school’s goals
• Your standards and local educational guidelines

Elements of an Effective AI Prompt

AI Models perform better when instructed to adopt an expert role. The role should be specific to your task, for example:

Curriculum Planning Role Prompt:  You are a curriculum designer and author with a focus on UbD and Bloom’s taxonomy.

Supporting Student Exam Prep Prompt:  Imagine you’re a test prep coach familiar with spaced repetition and retrieval practice.

Supporting Struggling Readers Prompt:  Assume the role of a literacy specialist experienced in phonics, comprehension strategies, and differentiated instruction.

All prompts should include a clear, specific description of your task. Think about how you would describe it to a colleague.

• Clearly describe the end result you are looking for
• Describe the impact you want to have on your audience
• Describe any steps or best practices the AI model should follow

The RISEN AI Prompt Framework

AI Prompt Frameworks are specific guides or techniques that help us quickly create high-quality prompts. You can use the RISEN prompt framework to help you craft an effective prompt. Remember:

• Be clear, specific, and conversational
• Consider your target audience
• If at first you don’t succeed, try again!

Example RISEN Prompt:

• Role: You are a DP History teacher designing a student-facing Source Evaluation Pack to build students’ ability to use primary sources in causation essays.
• Instruction: Create a downloadable resource pack in PDF form that a class can use in one 50–75 minute lesson or as a short homework sequence.
• Steps: Pick the topic case and gather three primary sources with provenance.
• Write one focused causation question.
• Build a single student page with provenance checks, short source tasks, and a model paragraph
• Add a brief rubric and short teacher notes
• End Goal: A student-ready pack for Year 2 DP History students who know the content well but need clearer guidance on analytical source use. It should help them move from recounting what sources say to judging provenance and using evidence purposefully in a causal argument (a skill directly tied to Paper 1/2 performance).
• Narrow: The pack must be ready to distribute to students with no extra teacher prep. Limit the pack to one topic case and three primary sources.

Final Thoughts

Remember, the best tasks for AI are tasks where you can easily identify a high-quality response, and tasks that you have relevant classroom context about (student performance, interests, needs, etc.)