Why Use Prompts with a Large Language Model Rather Than an AI Tool or App

Why Use Prompts with a Large Language Model Rather than a Dedicated AI Tool?

In the previous module, you were introduced to Kahnmigo, a tool that incorporates AI for dedicated purposes. This is a quick and easy way to get specific content. But what if you would like to create something the app isn't designed to create? Or maybe the tool or app doesn't give you the control to create what you want? Or in some cases, the tool requires you to may a monthly fee. In cases like this, you can use an LLM (Large Language Model like ChatGPT, Co-Pilot, Gemini, etc.) that you feed instructions to in order to get the output you want in the form of a written or even spoken conversation. This gives you more control over what the AI will create. These instructions that you give the the LLM, telling it you what it to create a called a prompt.

The Important of Learning to Write Good Prompts

Why is it Important to Learn to Write Good Prompts?

A prompt is the instructions you give to the large language model. The way you phrase a prompt significantly impacts the quality of what the AI creates for you. Just like giving clear instructions to a colleague or student, a well-crafted prompt ensures the AI understands your request and delivers useful results. Effective prompting can save time, reduce frustration, and unlock the full potential of AI tools in education. This module will help you to create effective prompts.

What is a Prompt Framework and How It Can Help You Write Better Prompts

An easy and effective approach to prompt writing is to use a prompt framework. Frameworks serve as templates or guides, helping users craft prompts that are clear, specific, and tailored to their goals.

Prompt frameworks are especially useful for educators, as they reduce the trial-and-error process and ensure consistent results when creating content, generating explanations, or solving teaching challenges. By following a framework, you can systematically approach any AI interaction and achieve better outcomes.

AI Large Language Model Options

You have many AI LLM options to try, such as Microsoft Copilot, which has been added to the college account for faculty and students. More capable options, including reasoning models and multimodal access, are becoming available.

The CRE Framework: An Example of a Prompt Framework

One commonly used framework is the CRE Framework. The CRE Framework is a simple yet powerful tool to guide your prompt creation. The acronym stands for Context, Request, and Example. By addressing these three components, you ensure your prompt provides the AI with enough information to generate high-quality results. Read below for details of the 3 parts of the CRE Framework. You will also find examples to provide clarity on how to build the best prompt.

Overview

Start with a podcast below discussing the CRE Framework.

Examine briefly the Study Guide created with NotebookLM, which produced the above podcast. Consider how these learning support contents (audio, self-assessment questions, terminologies, etc.) may help your students.

Context

Definition: Provide the AI with background information or the situation in which the task is taking place. This step sets the stage for a relevant and accurate response.

Why It Matters: AI systems rely on context to understand the scope and purpose of your request. Without it, the AI might generate generic or irrelevant responses.

Examples:

Poor: "Write a lesson plan."
Better: "Write a lesson plan for an introductory level technical college biology class."
Best: "Write a lesson plan for an introductory level technical college biology class on the topic of photosynthesis, incorporating a hands-on activity and group discussion.

AI Says: The first component of the CRE Framework is Context. Providing context means setting the stage for the AI by giving it background information relevant to the task at hand. This could include details about the subject matter, the intended audience, or any specific constraints that should be considered. By establishing a clear context, you help the AI understand the environment in which it is operating, which is crucial for generating appropriate and relevant responses.
Understanding the CRE Framework in Generative AI Prompting - visual selection.png
The infographic and text are from Napkin.ai.

Request

Definition: Clearly state what you want the AI to do. Be specific about the task, format, and level of detail you need.

Why It Matters: A vague or incomplete request can lead to responses that miss the mark. Specific requests ensure that the AI understands your expectations.

Examples:

Poor: "Make a quiz."
Better: "Create a 5-question multiple-choice quiz on the causes of World War II."
Best: "Create a 5-question multiple-choice quiz on the causes of World War II for college students. Include one question on each of the following: the Treaty of Versailles, economic factors, and political tensions."

AI Says: The second component is the Request. This is where you articulate what you want the AI to do. A well-defined request is specific and clear, guiding the AI on the expected outcome. Whether you are asking for a summary, a creative story, or a technical explanation, the request should be straightforward and unambiguous. This clarity helps the AI focus on delivering the desired result without unnecessary deviations.
Understanding the CRE Framework in Generative AI Prompting - visual selection (1).png
The infographic and text are from Napkin.ai.

Example

Definition: Provide a sample response or format you want the AI to follow. This acts as a guide for the AI to structure its output.

Why It Matters: Examples clarify the desired tone, structure, and level of complexity, making it more likely the AI will deliver exactly what you need.

Examples:

Without Example: "Summarize this article."
With Example: "Summarize this article in 3 bullet points, like this:
- Key point 1.
- Key point 2.
- Key point 3."

Tip: If you don't have an exact example, describe the format or tone you expect, such as "in plain language," "with examples," or "in an academic style."

AI Says: The final component of the CRE Framework is Example. Providing an example can significantly enhance the AI's understanding of your request. An example serves as a reference point, illustrating the format, tone, or style you are aiming for. By including an example, you reduce the likelihood of misinterpretation and increase the chances of receiving a response that aligns with your expectations.
Understanding the CRE Framework in Generative AI Prompting - visual selection (4).png
The infographic and text are from Napkin.ai.

Conclusion

The following infographic and text are from Napkin.ai.

In summary, the CRE Framework—Context, Request, and Example—serves as a robust guideline for creating effective prompts in Generative AI. By thoughtfully addressing each of these components, you empower the AI to generate high-quality results that meet your needs. Whether you are a novice or an experienced user, incorporating the CRE Framework into your prompting strategy can lead to more successful interactions with AI systems.
Understanding the CRE Framework in Generative AI Prompting - visual selection (3).png

The CRE Framework in Practice

Here’s how the CRE Framework might look in real scenarios:

Scenario 1

Writing Lesson Plans

Prompt:

Context: "You are an instructional designer creating content for community college science teachers."
Request: "Generate a 30-minute lesson plan on the water cycle."
Example: "Include sections for learning objectives, an engaging activity, and a summary assessment at the end."

Scenario 2

Explaining a Complex Concept

Prompt:

Context: "Explain mitosis to college students studying biology."
Request: "Write a short explanation using simple language and a step-by-step format."
Example:
- "Step 1: The cell prepares to divide.
- Step 2: Chromosomes line up in the middle.
- Step 3: Chromosomes are pulled apart."

Scenario 3

Creating Discussion Questions

Prompt:

Context: "You are a college instructor facilitating a discussion on climate change."
Request: "Write 5 open-ended questions that encourage critical thinking."
Example:
- "What are the main human activities contributing to climate change?
- How might climate change impact future generations differently than ours?"

For more information on the CRE Framework, Common Sense has created an online Course AI Foundations for Educators.

You can also download their repository with examples of several CRE prompts for education.

Tips for Applying the CRE Framework as an Educator

To get the most out of the CRE Framework:

Practice with familiar tasks. Start with a task you perform often, like generating activities or summarizing content.
Iterate and refine. Adjust the level of detail and examples in your prompt based on the quality of the AI's response.
Teach your students. Share the CRE Framework with students to help them interact effectively with AI tools.

By incorporating the CRE Framework into your workflow, you’ll create prompts that are purposeful, structured, and tailored to achieve your instructional goals.