IT Teaching Resources

Classroom Resources

AI Cases in Education

Leveraging generative AI tools to support teaching and learning

Article Technology Research

Presenter: Reuben Thiessen
Recording of the session:

Slides from the session

Central Questions:

  • What are some GenAI methodologies I can incorporate into my project?
  • How do I get started integrating these ideas into my process?
  • What are some considerations I should keep in mind when prompting ChatGPT?

Key Quotes:

“Perhaps you’re not immediately seeing how you use this in a learning or teaching use case, but as you start to work through it, that use case becomes more apparent.” (8:13, Reuben)

“One of the interesting things with a large language model is it can bring out interesting perspectives that you might not have originally thought of and new dimensions that you could explore in scenario building.” (8:50, Reuben)

“You could also generate data that could test your assumptions and kind of get you further along the road of user testing before you actually spend the time and expense with real humans.” (15:20, Reuben)

“[Prompt engineering is] turning into this whole art, and not necessarily all just about prompt engineering or getting the prompts right, but it really starts to make you think through the entire process of the hidden assumptions you might be coming in with and the need to be explicit about that.” (50:49, Reuben)

Take-aways:

  • Three D’s
    • Dreaming
      • Definition: Helping you think
      • Examples: Brainstorming, Concept expansion, Concept mapping, Collaborative Writing, Scenario building, Simulating discussions
    • Drudgery
      • Definition: Lightening your load
      • Examples: Summarization, Data Cleansing, Progress Tracking, Content Moderation, Synthetic Data, Review and Feedback
    • Design
      • Definition: Building your content
      • Examples: Lesson Plan Generation, Project Planning, Content Personalization, Accessibility Design, Interactive Experiences, Curriculum Mapping
  • Prompt Engineering
    • ChatGPT has a tendency to latch onto surface level features
      • Add more specificity like concepts, names of researchers, books, and other scaffolding
      • Give ChatGPT some assumptions about prior knowledge, such as knowing about a teaching model
        • Define any concepts that it may have fuzzy understanding on
    • ChatGPT gives a good starting point but going deeper can be challenging
      • Ask for ChatGPT to include other concepts like engagement or practices
      • Give ChatGPT more context like using it in a classroom or with 5th graders
  • Resources
  • Community
  • Materials linked in the session