AI Teaching Exchange

Willamette University

Author

Lucas Cordova

Published

August 5, 2025

Introduction

Why Customize AI for Learning?

  • Off-the-shelf chatbots ≠ pedagogically aligned
  • Customized tools allow:
    • Targeted scaffolding
    • Domain-appropriate reasoning
    • Support for diverse learning objectives

Learning Objectives Supported

  • Critical Thinking
  • Content Knowledge Acquisition
  • Metacognition & Self-Regulation
  • Process-based Reasoning
  • Confidence Building through Conversation

Research: Pedalogical

Structured AI for Deep Learning

  • ZPD: Prompts target the “next step” of understanding
  • Cognitive Load Theory: Structured sequences reduce overload
  • Constructivism: Learners build knowledge by explaining

Chatbot Roles We Are Exploring

  • Peer: Encourages openness, low-stakes exploration
  • Tutor: Offers step-by-step guidance, structured tasks
  • Instructor: May be too authoritative, risks overreliance

Key Insight: Role framing matters. Pedagogical alignment enhances learning impact.

Summer Learning Circle

  • Faculty-led exploration of AI use in courses
  • Focused on:
    • Chatbot customization (OpenAI, Claude, etc.)
    • Prompt engineering
    • Canvas LMS integration
  • Outcome:
    • Canvas shell toolkit for Willamette faculty

Use Case

Scenario

Course: Introductory Biology

  • Student feels stuck: “I don’t know where to start.”
  • Assignment: Interpret data from photosynthesis lab

Assignment Details

Course: Intro Bio
Topic: Data Interpretation
Objective: Make a claim & justify it with experimental data

Prompt:
“Using the data from our photosynthesis lab, make a claim about how light intensity affects oxygen production in Elodea.”

AI Conversation Flow

title: AI Conversation Flow (Feynman Style) format: revealjs layout: false

Student: I’m not sure what to write.

Student: The oxygen levels changed with light.

Student: I think plants use light to make energy, and that makes oxygen?

Student: Maybe… more light increases the rate of photosynthesis?

Chatbot: No worries — let’s try explaining what’s going on in your own words.

Chatbot: Can you explain why light might affect oxygen levels — as if you were teaching it to a friend who missed class?

Chatbot: Yes! That’s photosynthesis. So if more light means more oxygen, what’s the connection?

Chatbot: That’s a solid claim! Now, how could you support that using the data from your lab?

Why It Works

  • Helps with initiation, not just revision
  • Low-stakes reasoning partner
  • Reinforces scientific process thinking
  • Scalable to other lab-based courses

Closing & Invitation

  • Custom AI tools let us align pedagogy + technology
  • They support learner confidence and depth of thought
  • Join us in:
    • Using the Canvas shell
    • Designing your own assignment-aligned tools
    • Reimagining AI not as automation—but as amplification of good teaching
    • Reaching out if you want to try a customized AI tool in your course!

References

  • Vygotsky, L. S. (1978). Mind in Society: The Development of Higher Psychological Processes. Harvard University Press.
  • Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257–285.
  • Bruner, J. (1961). The act of discovery. Harvard Educational Review, 31(1), 21–32.
  • Chi, M. T. H., et al. (1994). Eliciting self-explanations improves understanding. Cognitive Science, 18(3), 439–477.
  • Clark, R. C., & Mayer, R. E. (2016). E-learning and the science of instruction: Proven guidelines for consumers and designers of multimedia learning. Wiley.
  • Cordova, L. P., Sanders, A. L., Mendoza, T. J., & Walia, G. S. (2025). Pedalogical: Feedback tool to reduce software vulnerabilities in non-security computer science courses. FIE 2025 (Accepted).