a differentiation solution?
Condensing text has always struck me as one of gen-AI’s genuine strengths—especially with passages only a page or two long. Because colleagues and I constantly wrestle with teaching complex ideas to readers at wildly different levels, I decided to run a little experiment.
I grabbed a section from an open Canadian-history textbook on Winnipeg’s water supply and its century-long impact on Shoal Lake First Nation. (Copyright dodged!) Then I sent the same passage through two “grade-five level” text-levelling tools. After the fun I had last week coding responses (sadly I am not being sarcastic) I did a bit of the same here. The results were fascinating. My hunch is that these tools perform better in tightly structured subjects like science or math, but I wanted to see how they’d handle a topic that matters deeply in Winnipeg and which structures of power and colonial legacy have significant impact.
In a perfect world you’d use an AI system that lets you spell out the key concepts that must survive the rewrite, but that raises the stakes for prompt quality. For this assignment I stuck with true paste-and-go tools—the kind that lure in brand-new or still-skeptical AI users.
I’ve bundled my heuristic, the side-by-side outputs, and a brief analysis in a Genially presentation (link below). Make sure to use the show interactive elements button in the top right corner, so that you don’t miss any interactive content. I’d love to hear your thoughts.
