34 Thinking with AI
Ashutosh Mohandas Kotian, Postgraduate Student, Sustainable Food Systems and Food Security, United Kingdom
During my postgraduate studies in sustainability, I was required to write a critical essay evaluating whether carbon offset markets effectively protect biodiversity. I had gathered a wide range of literature and developed key points, but I struggled to organise these ideas into a coherent academic argument. This challenge, how to structure competing perspectives while meeting academic standards, prompted me to explore the use of GenAI as a support tool for learning, particularly in the early planning and structuring phase of the assignment.
In accordance with my University’s guidelines, I approached GenAI use not as a way to generate assessed content, but as a reflective and exploratory aid to enhance my understanding of argumentation and improve my academic writing process.
I used ChatGPT (GPT-4) to support the ideation and structure-planning stage of my essay. My initial prompt was:
“What are three different ways to structure an essay that argues carbon offset markets fail to protect biodiversity?”
This generated useful framing ideas, including thematic, chronological, and stakeholder-centred structures. These responses helped me visualise the strengths and limitations of each approach and explore unfamiliar perspectives. I critically evaluated the suggestions, selecting elements that aligned with my intended thesis and the assignment’s learning outcomes.
I then deepened the interaction by prompting:
“What weaknesses might this structure have in terms of coherence or clarity?”
“What counterarguments would challenge this approach?”
These responses encouraged me to anticipate critiques and refine my argument accordingly. Importantly, I did not use any GenAI-generated content in my submission. The AI outputs served only as scaffolding for my thinking. I rephrased, reorganised, and reanalysed the ideas independently, ensuring all submitted work was my own original writing and critical reasoning.
In line with good academic practice, I also reflected on the quality of the GenAI’s responses. When outputs were superficial or unclear, I adjusted my prompts by adding more context (e.g., specifying audience, word count, and focus). This iterative process taught me how to engage more purposefully with GenAI and recognise its limitations.
Using GenAI in this way strengthened my ability to structure arguments, evaluate reasoning strategies, and improve the clarity of my academic writing. The process deepened my understanding of how structure influences argument strength and helped me make intentional decisions about the flow and logic of my essay.
Engaging in dialogue with GenAI helped me to slow down, test alternatives, and think critically about my framing choices
This use of AI also supported metacognitive learning. I became more aware of my habits as a writer, particularly my tendency to overexplain or under-prioritise key points. Engaging in dialogue with GenAI helped me to slow down, test alternatives, and think critically about my framing choices. I wasn’t passively accepting responses but actively refining and challenging them.
The experience also enhanced my confidence. I found that I was more willing to experiment with ideas and restructure drafts early on, which made the overall writing process less intimidating and more iterative. Rather than rushing to a conclusion, I learned to value exploration and refinement.
Overall, GenAI acted as a thinking partner, not a content creator. This approach aligns with academic integrity and the values promoted by my university it is enabling active learning, personal reflection, and improved independence in my academic work.
GenAI Tool(s) Used:
ChatGPT (GPT-4) via ChatGPT Plus subscription
Occasionally Claude.ai for alternate perspectives (non-assessed use only)
