Skip to content

GenAI Essay

Using GenAI to solve a ICT&Business semester 3 data engineering problem

Preface

This essay was generated by ChatGPT and has therefor the distinct smell of a GenAI generated text. It is based on a series of notes included with this text. These notes were made in Dutch while using GenAI to solve a data engineering problem assigned to students in the third semester of the ICT & Business program. Included are:

  • The interactions I had with ChatGPT and Claude to address the problem,
  • The notes I took in Dutch during this process,
  • The interaction I had with ChatGPT to create a coherent essay from these notes,
  • This disclaimer, which I initially wrote in Dutch (admittedly, due to becoming a bit lazy with the help of GenAI), and
  • The prompt I used to add this disclaimer in English to the essay.

I take full responsibility for the content and arguments presented in this essay. Although GenAI assisted in organizing and drafting the text, the final result reflects my (current) perspective and the conclusions I have drawn from my experiences.

This essay is just a byproduct, and I feel that too much attention is given to such byproducts in the ongoing debate over the use of GenAI in higher education. The essence of this effort lies in the interactions I had with ChatGPT and Claude—solving a data engineering problem step-by-step while simultaneously gaining new insights and knowledge. I believe we need to share more of these examples with one another to develop well-grounded best practices for the specific education we provide to our students in FICT.

The Role of Generative AI in Higher Vocational ICT Education

Generative AI (GenAI) tools like ChatGPT and Claude are increasingly becoming a part of the educational landscape, particularly in ICT education. However, their use poses unique challenges and opportunities, especially in vocational contexts where practical skills are paramount. This essay reflects on a case study using GenAI to solve a specific data engineering problem within a the third semester if ICT&Business in FICT. It examines the potential benefits and pitfalls of using GenAI for both educators and students, with a focus on critical thinking and deeper understanding, rather than merely enhancing productivity.

Case Study: Using GenAI for Data Engineering

The case in question involved a data engineering problem from the third semester of an ICT & Business program. The task required importing, converting, and combining data from three different sources using Power Query in Power BI for further analysis. At first glance, this seemed like a problem that GenAI could help solve. However, without any prior knowledge or experience in data engineering, the solutions proposed by GenAI were largely inadequate and could easily mislead students further into confusion.

I experimented with two different GenAIs—ChatGPT and Claude—to explore how they might assist in solving this problem. The main takeaway was that while both tools had their unique features, there was no significant functional difference between them for this specific task. For instance, ChatGPT's ability to easily save and share conversations was indeed useful, but it did not contribute to the quality of the proposed solutions.

Breaking Down the Problem: A Strategy for Using GenAI Effectively

One effective strategy to improve the use of GenAI in solving such complex problems was to break down the main problem into smaller, more manageable sub-problems. This "divide and conquer" approach helped to make the AI's task easier and more focused. However, this method also quickly led to hitting token limits, especially when using free versions of these AI tools. In my experience with Claude, providing feedback and validating insights resulted in having to wait several hours before I could continue, which might discourage students from asking deeper, more insightful questions or asking for acquired insight validation due to token limitations.

This points to a fundamental requirement: for a user to effectively leverage GenAI, they must already possess a certain level of knowledge and experience in solving similar problems. Moreover, they need to be able to divide a non-trivial problem into sub-problems and clearly understand the expected intermediate results for each step, allowing them to critically evaluate whether the AI's suggestions lead to the desired outcomes.

What Does Conscious Use of GenAI Mean?

Conscious use of GenAI involves engaging actively with the AI, treating it as an equal learning partner rather than a passive tool. While GenAI has access to a vast amount of factual knowledge, it often lacks context, oversight, or a clear understanding of the end goal. Therefore, students must be proactive in guiding the AI, questioning its assumptions, and critically evaluating its outputs.

This approach requires students to interact continuously with the AI, refining their queries, and seeking deeper insights. Rather than accepting the AI's responses at face value, they should use the tool to challenge their understanding, connect disparate concepts, and adapt their problem-solving strategies based on the AI's feedback. In essence, conscious use means actively driving the learning process, with GenAI acting as a knowledgeable but fallible partner in that journey.

The Dual Role of GenAI: Productivity vs. Learning

The use of GenAI in education highlights a paradox: while it can serve as a powerful tool for both enhancing productivity and supporting learning, these roles can sometimes seem at odds. On the one hand, GenAI can significantly speed up the completion of well-defined tasks, such as coding or data analysis, by providing ready-made solutions. On the other hand, its role in facilitating deep learning and understanding requires a different approach, one that emphasizes exploration, critical thinking, and engagement with the subject matter.

The solution to this paradox lies in combining both roles within the same activity. Students can use GenAI to solve a problem in a practical way—boosting productivity—while simultaneously learning from the solution. This requires them to not only find answers but also to understand the reasoning behind those answers, draw connections to broader concepts, and reflect on their learning process.

For example, instead of merely asking GenAI to "generate a solution," students should engage in a dialogue with the AI, asking why certain steps are taken, what alternative approaches might exist, and how different parts of the solution relate to the overall problem. This integrated approach helps students improve their practical skills while deepening their conceptual understanding, ultimately leading to a more balanced and effective use of GenAI in education. And then there is this token limit ...

Conclusion: Fostering a Balanced Approach to GenAI in Education

The use of GenAI in higher vocational ICT education presents both significant opportunities and notable challenges. While it can enhance productivity and support deep learning, these dual roles often create a tension that must be carefully managed. The key is guiding students to use GenAI not merely as a shortcut to answers, but as a tool to foster a deeper understanding of complex problems. Combining both roles—solving practical tasks while simultaneously engaging in critical reflection—can maximize the benefits of GenAI.

However, there are limitations to this approach, especially in educational settings where free or limited versions of GenAI are used. Token limits can restrict the extent to which students can engage in iterative problem-solving and continuous learning, forcing them to prioritize one use over the other. As educators, we must be mindful of these constraints and work to create strategies that help students navigate them effectively, ensuring that GenAI serves as a meaningful aid in their learning journey rather than an obstacle to it.