Navigating the Impact of GPT on Student Presentations: A Teacher's Dilemma.

Teaching in the Age of GPT: Reflections on Classroom Experiences

Well, we are very confused about teaching in the time of GPT. This past week has been evolutionary for me. In the next three blog posts, I will discuss three classroom cases that illustrate my experiences, my students’ evolving attitudes, and reflections on the best strategies for using GPT as a tool for learning.

Dictionary:

1. evolutionary: involving a gradual process of change and development:

   The change has been evolutionary rather than revolutionary.

2. revolutionary: completely new and having a great effect:

   Penicillin was a revolutionary drug.

Case 1: The Presentation Paradox

I recently challenged my students on their use of GPT to develop presentations. The assignment was part of a research project for Global Citizenship, focusing on community issues. One group chose to explore food waste on a global scale, comparing two countries.

Assessment Criteria

To ensure high-quality work, I provided a detailed assessment framework. Students were free to select their case study countries and had to deliver a presentation that met the following criteria:

Section	Task
1. Introduction	Define food waste and its importance. Briefly introduce two case studies. State your argument.
2. Case Study 1: France	Describe France’s law banning supermarkets from throwing away food. Explain how it helps reduce waste and hunger. Discuss challenges (e.g., enforcement, costs).
3. Case Study 2: India	Explain how poor storage and transport cause food waste in India. Discuss impact on farmers and consumers. Compare with France’s approach.
4. Evaluation	Compare both case studies. Discuss effectiveness and challenges. Suggest a possible solution combining both approaches.
5. Conclusion	Summarize key points. Restate the importance of reducing food waste.
6. References	Each case, you need 2 different information sources: e.g. an interview, a video, an article, Wikipedia entry etc. All must be referenced by APA, and then explain why you used that source with CRAAP (1-2 sentences). APA: https://www.citationmachine.net/

Their presentations were, without a doubt, impressive. The structure was impeccable, the design was professional, and their delivery was confident. It was clear that they are highly skilled in presenting information. However, there was one glaring issue: GPT’s tendency to provide the same general information unless specifically prompted otherwise.

The Role of GPT in Their Research

I emphasized the importance of thorough research and proper referencing in this task. GPT had done this part for them seamlessly. With the help of citation tools, they produced flawless citations. At first glance, everything seemed perfect.

However, upon closer inspection, a critical flaw became evident. One group, for instance, used Singapore as a case study. Their citations referenced official reports from Singapore’s waste management agency—up-to-date and neatly presented. But the problem? GPT-generated content lacked the depth of analysis needed to differentiate between general overviews and the nuanced, specific strategies employed in each country.

The Grading Dilemma

Despite their strong presentations, not a single student received over 70%—which, understandably, frustrated them. Their main argument? How can a teacher who uses GPT to design lessons penalize students for using GPT to create presentations?

And honestly? They had a point.

My conclusion: GPT is not as thorough as I expected and will crumble upon the pressure of a detailed assessment with a focus on research quality and students' understanding.

Case 2: Rethinking Research in Coastal Management

My Year 12 students were wrapping up their coastal management unit with final presentations. The task was clear: compare two different locations, analyze their coastal strategies, define success criteria, evaluate the effectiveness of the strategies, and compare both case studies. Before this, we had studied Mumbai and explored various hard and soft engineering strategies around the world.

To ensure quality research, I emphasized the importance of credible sources, modeled a CRAAP analysis, and demonstrated how GPT often reduces complex analysis to generic statements. However, I allowed them to use GPT freely, provided they supported their findings with relevant data.

1. Introduction	Define key terms (coastal management, success, developing/emerging countries). Briefly introduce the strategies you will analyze. State your argument (to what extent you agree).
2. Case Study 1:	Point: Introduce the strategy and location. Evidence: Provide key data on success and failures. Explanation: Analyze effectiveness (economic, social, environmental impact). Link: Compare with another strategy.
3. Case Study 2:	Point: Introduce the strategy and location. Evidence: Provide data on success and failures. Explanation: Assess effectiveness and compare with hard engineering. Link: Relate to another approach.
4. Evaluation: Compare Strategies	Compare short-term vs. long-term success. Weigh economic vs. environmental sustainability. Assess whether any strategy works universally or needs adaptation.
5. Future Solutions & Improvements	Suggest a balanced approach (integrating different strategies). Propose policy recommendations for better coastal management.
6. Conclusion	Summarize key findings. Restate your overall judgment. End with a strong final statement.

The result? Four mini-groups, four presentations, four hours of classwork over Friday and Tuesday. What I witnessed was remarkable:

• I had never seen this group so focused, motivated, and engaged in research.

• They treated GPT as a research assistant rather than a content generator.

• They formulated hypotheses, verified them with GPT, then conducted deeper research.

• They collaborated effectively, refining and structuring their findings.

  
  

The final presentations were impressive. As a class, we concluded with a “feedforward” discussion in the forum, allowing for constructive feedback and future improvement strategies.

My conclusion: GPT can be used as a powerful research assistant tool if we are able to motivate students to use it that way and bring feelings (pride, satisfaction, state of flow, exploration) to the classroom (and surely, a detailed assessment will help).

Case 3: The GPT-Only Presentation vs. Research-Based Learning

Year 13 Geography. Exam Unit 4. A 60-mark essay on food security. To improve retention, I decided on a case-study approach, where each case would be designed and presented over the course of a few days.

Step	Description	Key Questions
1. Introduce a Strategy	Provide an overview of the strategy, including its purpose and implementation.	What is the strategy? Where and when was it implemented? Why was it introduced? How does it work?
2. Define "Success" Clearly	Explain what makes the strategy successful in its specific context.	Economic Success: Does it increase farmers’ incomes, reduce food prices, or boost national agricultural productivity? Social Success: Does it reduce hunger, improve nutrition, and benefit small-scale farmers? Environmental Success: Is it sustainable? Does it protect biodiversity, soil health, and water resources?
3. Evaluate Different Perspectives	Analyze the strategy’s effectiveness from multiple angles.	Short-Term vs. Long-Term Success: Did it work initially but cause long-term problems (e.g., soil degradation in the Green Revolution)? Who Benefits?: Did large agribusinesses profit while small farmers struggled? Challenges & Limitations: Was success dependent on specific climate, political, or economic conditions?
4. Use Two Credible Sources per Case	Support findings with diverse sources and justify their credibility.	Choose at least two sources (e.g., interview, video, article, Wikipedia entry). Cite them in APA format and explain why each source was used.

The first case was GM crops in Brazil. I joined my students (only two in the class) in preparing a presentation—mine based on FAO and World Bank sources, theirs entirely produced by GPT. Confidently, they presented their slides, while I had done thorough research, curated relevant visuals, and critically analyzed the topic.

What happened next was enlightening. Instead of giving immediate feedback, I:

1. Presented my version for comparison.

2. Reopened their presentation and asked probing questions:

   • Where does this conclusion come from?

   • Where is the supporting data?

   • How do you relate these concepts?

   • How did you measure success?

   • How do you know this is true?

3. Revealed my GPT chat history—showing how their content was nearly identical to what GPT had produced for me in seconds.

   
   

They scored 50% and were actually grateful for it. The following day, I asked them to use my structure, visuals, and research to redo the Brazil presentation.

Case 4: The Shift to Research-Backed Work

Expecting a redo of Brazil, I was surprised when they chose India instead. What emerged was a refined approach:

• I provided citations, structured arguments, and supported conclusions with visuals.

• They integrated GPT’s organizational structure while improving paragraph flow and coherence.

  
  

The result? A polished, well-researched presentation that scored 85%—falling short only in underdeveloped sections on limitations and future solutions. In just 24 hours, their work had transformed.

When I asked them about their learning process, they reflected:

• “We researched many sources.”

• “We used different reports.”

• “We used visuals to support our ideas.”

• “It’s easier to explain processes with graphs.”

My conclusion: GPT provides a neat and effective structure for work. Demand your students to dive deep with other sources (by design of assessment, force them to explore alternatives) and require them to support their claims with visuals, expert opinions etc.

Final Thoughts (by GPT)

These four cases illustrate a profound shift in the teaching-learning dynamic when integrating GPT into the classroom. Initially, students relied on GPT as a shortcut, producing polished yet shallow work that lacked critical analysis and depth.

However, as expectations for credible sourcing, original thought, and data-backed arguments were reinforced, their use of AI evolved. Instead of replacing research, GPT became a research assistant—helping to refine ideas, verify assumptions, and structure content effectively.

The turning point came when students realized that while AI-generated content can be a strong starting point, it lacks the nuance, contextual understanding, and critical evaluation that true academic work requires.

For educators, the key takeaway is that AI should not be dismissed or banned but guided towards meaningful application. By setting clear expectations, modeling effective use, and emphasizing source evaluation, teachers can turn GPT from a tool of convenience into a catalyst for deeper learning.

The cases show that when students are challenged to think critically beyond AI-generated responses, they engage more, take ownership of their work, and develop essential research and analytical skills. In this evolving landscape, the role of the teacher is no longer just to assess knowledge but to train students to navigate, question, and enhance AI-driven insights—skills that will be invaluable in their academic and professional futures.

And how can you NOT love GPT?