Several professors from the University of Munich were on stage discussing a problem that universities everywhere are now facing: students are increasingly writing essays with the help of generative AI. And, honestly, many professors seem genuinely uncertain about what to do next, because the essays often turn out to be remarkably good.
Then one of the professors said something very interesting:
“We all understand that generative AI almost never produces a truly strong essay from the very first prompt. Which means the student somehow arrived at that final version: refining ideas, restructuring arguments, clarifying points, improving depth and coherence. So why not ask students, during the defense of their essays, to also submit the history of prompts they used throughout the process?”
And that thought struck me as genuinely profound.
Because when someone is truly working with AI — rather than simply trying to get a finished result from a single prompt — genuinely good writing almost always requires a huge number of iterations. Sometimes I personally spend an entire day on one serious writing task simply because of the constant refining, questioning, restructuring and searching for more precise formulations.
And if you later rewind the entire process and look through the chain of prompts, something very interesting begins to emerge: you can actually see the person’s thinking process unfolding in real time — what they notice, where they hesitate, how they deepen an idea, which weaknesses they recognize in the text, and how capable they are of thinking independently in the first place.
And perhaps this will eventually become one of the clearest ways to distinguish people who use AI to amplify their own thinking from those who simply try to outsource thinking to the machine.
📱 Mykola Latansky. Subscribe!
Leave a reply