Designing a curriculum that requires students to use AI critically and ethically (Part 1)

Can we design a task where students cannot cut and paste their responses?

I am reflecting on a meeting on using AI in education and asking GPT to convert my reflection into actionable knowledge. Here is how it goes.

 User

You said:
"1. Task designed should be broad enough that requires co-thinking with AI by (a) assigning GPT with a role, (b) prompting GPT to solve the problem (c) reflecting about their prompting process. 2. For a task to be broad and of significance to the students, Jonathan suggested "better food in the canteen" as a good example. 3. In (a) Assigning GPT with a role can be used to (i) provide perspectives (e.g in canteen project, it can be the perspective of teachers, students and vendors. And Better food in the canteen may be in the form of better preparation, better taste, healthier choice etc) 4. In (b) - when prompting and interacting with GPT, students need to be critical and demonstrate higher-order thinking (analyse and evaluate what GPT said part by part to bring about new insight) 5. In (c) - (i) reflect through the interaction with GPT, how GPT cannot do - e.g to empathize the situation in context (e.g how the canteen vendor feels about having a new dish) and the need to conduct primary research. (ii) How students interpret GPT is a research gap based on my knowledge. If students interpret GPT information at the surface level, it is unlikely to provide any insights at the end of the conversation, and we can help students frame the interpretation by comparing it with our knowledge, what GPT said earlier (if it contradicts itself), and with other AI tools. We can look deeper into the "co-thinking" framework when interacting with GPT with 4 steps: Role, Context, Sequence, and Structure. Dr Andrew introduced us to https://leonfurze.com/blog/, a PHD student reflecting on AI in Education. Will be a good person to follow.

1. Designing the Task for Co-Thinking with AI

The curriculum should frame the task broadly enough to encourage co-thinking with AI, incorporating the following three steps:

  • (a) Assign GPT a role to generate diverse perspectives.

  • (b) Prompt GPT to solve the problem while critically analyzing responses.

  • (c) Reflect on the prompting process to identify AI’s limitations and the need for human-driven inquiry.

2. Framing the Problem with Real-World Relevance

To make the task meaningful and relevant to students, we select "Better Food in the Canteen" as the central theme. The problem can be explored in different ways:

  • Perspective-Based Inquiry: Students consider perspectives from teachers, students, and vendors.

  • Defining 'Better Food': Investigate aspects such as better preparation, taste, health benefits, and sustainability.

3. Assigning GPT a Role (Perspective-Taking for Diverse Insights)

When assigning GPT a role, students can:

  • Ask GPT to respond as different stakeholders (e.g., a canteen vendor, nutritionist, school administrator, or fellow student).

  • Explore constraints and opportunities for each stakeholder.

  • Generate potential policy recommendations based on stakeholder needs.

Example Prompts:

  • "As a school canteen vendor, what challenges do you face in preparing healthier meals?"

  • "As a student, what food improvements would make your meals more enjoyable and nutritious?"

4. Critical Analysis and Higher-Order Thinking in Prompting GPT

Students need to analyze and evaluate GPT’s responses to generate new insights by:

  • Comparing AI-generated ideas to personal knowledge and prior research.

  • Questioning contradictions within GPT’s responses.

  • Breaking down AI-generated ideas into actionable solutions.

  • Validating insights with external sources and primary research.

Example Reflection Questions:

  • Did GPT provide a feasible solution? Why or why not?

  • Does GPT contradict itself in different responses?

  • What further research is needed to verify AI’s suggestions?

5. Reflection on the AI Interaction Process

Students should reflect on:

(i) AI’s Limitations in Contextual Understanding

  • GPT cannot empathize with real-world challenges (e.g., how a vendor feels about financial constraints when introducing new dishes).

  • Primary research is necessary (e.g., interviews with canteen vendors or surveys with students) to supplement AI insights.

(ii) Developing Research Skills through AI Interpretation

  • If students interpret GPT's responses superficially, insights will be limited.

  • Encouraging students to compare AI-generated content with personal experience, prior knowledge, and alternative AI tools.

Example Reflection Questions:

  • What is missing in GPT’s responses that we need to find through primary research?

  • How does GPT’s response compare to real-life perspectives?

  • Are there inconsistencies in GPT’s responses, and how do they impact our understanding?

6. Structuring the Co-Thinking Framework (Role, Context, Sequence, Structure)

A structured framework for AI-augmented thinking:

  • Role: Assign GPT stakeholder perspectives.

  • Context: Frame the problem broadly but within a real-world scenario.

  • Sequence: Guide students through questioning, analysis, and iteration.

  • Structure: Encourage reflections that integrate AI insights with primary research.

7. Next Steps and Implementation

  • Develop lesson plans integrating AI co-thinking activities.

  • Create a structured reflection guide to help students critically analyze AI-generated responses.

  • Encourage primary research activities such as surveying students, interviewing vendors, and testing menu changes.

  • Follow researchers such as Leon Furze for insights on AI in education.

This curriculum leverages AI as a thinking partner rather than a replacement for student inquiry, ensuring deeper engagement with real-world problem-solving in design and technology education.

Comments

Popular posts from this blog

2024 Thrive 3 - Self-evaluation of sketching using pen

My reflection with AI about the STEM project