This morning, I came across a Medium article titled "I'm a Professional Editor, and These Phrases Tell Me You Used ChatGPT" (link). It pointed out red-flag phrases like "in the world of," "consequently," and "furthermore" - words that supposedly signal a piece was written by AI. But as someone who works with AI daily, I couldn't help but think: it's not the tool that's flat - it's the prompt.
AI writing is only as good as the prompt it's given. If you feed it vague, generic instructions, you'll get vague, generic prose. But if your prompt includes specific details, tone, structure, and intention, the output can surprise you in all the right ways.
Let's Look at a Real Example
One of our students - an incredibly gifted young coder and math enthusiast - once submitted an essay that opened like this:
Before (Telling): "I've loved math for as long as I can remember. It's always come naturally to me, and I enjoy solving problems. Math has helped me think logically and develop strong problem-solving skills."
There's nothing wrong with this paragraph. But it tells us how the student feels about math rather than letting us see it ourselves. Now, let's consider what happens when we revise it using a prompt that encourages storytelling and specificity. Same student, new prompt: "Write about how you discovered your love of math through moments in childhood. Include specific memories or experiences that shaped your relationship with numbers."
After (Showing): "When I was four, I surprised my mom by reading the clock correctly in the kitchen - 'It's 3:45,' I said, pointing proudly. She laughed, called my aunt, and said, 'He's doing fractions now?' Soon after, I started calculating change at the supermarket, holding coins in one hand and subtracting totals in my head. Every time I got it right, I felt like I had a superpower. Math wasn't just a subject - it was a language I spoke before I even knew how to write my name in cursive."
Suddenly, we're in the story with him. We can see the curiosity, feel the pride, and believe in his love for math - not because he said so, but because the details speak for themselves.
Why This Matters for Prompt Engineering
The takeaway isn't just about writing essays - it's about designing better prompts. Whether you're using AI to write, code, research, or brainstorm, the results hinge on how much substance you give it to work with. Asking for an "essay about loving math" will give you a cookie-cutter answer. Asking for a story about how and why that love began - prompting for moments, not summaries - unlocks a completely different tier of output.
So no, it's not fair to dismiss a piece just because it uses words like consequently or in the world of. Those phrases don't make writing robotic - lack of specificity does. The real magic lies in the details you feed the model.
If we want AI to "show, not tell," we need to prompt it the same way we'd guide a student or a collaborator: ask for experiences, not just conclusions.