Live More new Guides on the way! Get notified by signing up to CAST's newsletter..

Home > Guides

Creating course materials using ChatGPT

Contributed by

MadLab is a grassroots innovation organisation. We work with technology and people to shape our world for the better – combining a deep material understanding of new and emerging technologies with the tools and methods of participatory design.

Use this guide to understand how generative AI tool ChatGPT can help you plan and prepare course content for your charity. It talks about how you can use ChatGPT as a ‘co-pilot’ to develop course outlines and materials, including learning objectives, examples, and video scripts.

Jump to steps
Key software:
Notion, ChatGPT
How they set it up:
Followed simple on screen instructions
Setup costs:
None
Ongoing costs:
None
Guide in use since:
May 2024

Steps to creating course materials using ChatGPT

Tools like ChatGPT, Claude, and Google Gemini are known as generative artificial intelligence (GenAI). They use algorithms known as large language models (LLMs) to identify patterns in huge datasets, and then generate new text using these patterns, a bit like the predictive text tool on our phones.

Look at your current process when you are making a new course and think about where ChatGPT might help you to save time.

  • What steps do you normally follow when you are making a new course?

  • What types of content do you create regularly?

  • Which tasks take up a lot of your time?

  • Where do you sometimes need fresh ideas?

MadLab is a grassroots innovation organisation based in Manchester. They run lots of different projects, including courses in digital skills for young people.

They are a very small team, and they wanted to speed up how they made learning materials and marketing content. They also wanted to understand how to use ChatGPT so that they could share this knowledge with the people on their courses.

ChatGPT takes the form of a text conversation. You can ask ChatGPT questions to help understand your own thoughts, as if you were talking to a person. You can also give ChatGPT prompts or instructions and ask it to generate particular pieces of text or include specific information.

Try different ways of interacting with ChatGPT to see what works best for what you need. For example, you could try:

  • Open-ended questions for brainstorming

  • Specific, detailed prompts

  • Iterative prompts which build on each other

  • Asking for explanations and analogies

  • Telling ChatGPT to write in a ‘beginner friendly’ style, or for a particular reading level

Be creative! Explore what kinds of questions and prompts deliver the results that you want.

As you go along, make a note of what works and what doesn’t.

MadLab used ChatGPT to help create courses about programming languages, such as Javascript. They used different types of questions and prompts, for example:

  • Open-ended questions, like “What are the main topics to include in a beginners’ course about Javascript?”

  • Iterative prompts which build on each other, like “Using the list of topics you just gave me, write three key learning objectives for each topic”

  • Specific, detailed prompts, like “Using this summary, create a video script with a speaker pacing of about 100 words per minute. The tone should be conversational, and the language must be accessible to a non-technical audience, targeting a Flesch-Kincaid readability score of 9-12.”

  • Adding relevant information, for example “...In your answer please include the following points:...”, “Be sure to clearly define technical terms”

  • Asking for analogies, such as “Give me a useful analogy which helps to explain this programming concept”

  • Building on the results of previous prompts with new prompts.

MadLab found that being very specific in their questions helped ChatGPT to produce more useful content.

They also asked ChatGPT about the prompts themselves in order to get better answers. For example: “What additional information would be useful to include in the following prompt: ‘What are the main topics to include in a beginners’ course about Javascript?’”

Generative AI tools often provide inaccurate information. These are known as ‘hallucinations’. Check all of ChatGPT’s answers against trusted knowledge sources, including your own insight and experience.

MadLab could tell straight away when something wasn’t right because they are very familiar with Javascript.

Sometimes they asked ChatGPT to explain how it arrived at the answer it gave. AI tools don’t actually think or understand the words they generate the way we do, but it can still be helpful to see how an answer was produced. This approach allowed MadLab to adjust the prompts and instructions they used to get better results.

Keep track of how you find using ChatGPT:

  • Make a note of useful questions and prompts

  • Note which types of tasks ChatGPT handles well

  • Record any problems or issues you encounter

  • Review and update your notes as you explore

  • Read up on best practice to improve your prompting

Use this information to create a standard set of prompts for generating course outlines and materials. This might include:

  1. A prompt for overall course topics to develop an outline

  2. Prompts for breaking down each topic into subtopics, with examples

  3. Requests for learning objectives for each section

  4. Prompts for suggested tasks, activities, and assessments to deliver the learning objectives

  5. Prompts to generate text for slides, handouts, and video scripts

Make sure your new process includes regular points where you or a colleague will check that everything is accurate and appropriate.

MadLab found that ChatGPT was a useful sounding board to help them refine their ideas, and could help them create course content more quickly. However it did take a lot of time to find and develop the most useful prompts and instructions. They have turned what they found into a process which they can repeat.

Generative AI is a new technology. Lots of people are exploring how to use it safely and effectively. Think about sharing what you learn with your colleagues, and with other charities.

MadLab used Notion to share their new process for creating course content with their wider team. They are also exploring ways to share what they have learned with the young people on their courses.

Further information