All over the internet we hear the same worry from knowledge workers and creatives: AI will replace us, or at least flood the world with bland, machine-made slop.
I understand the concern. But the more I use AI in my own writing, the more I think differently. It isn’t putting me out of work. It’s actually helping me reach a degree of clarity and invention I might not have reached before. In some ways, it’s even making me better.
The key is interrogation. You can’t take the first answer and move on. The first pass from ChatGPT is often competent but generic. It’s polite, neutral, and a bit bureaucratic. And if you stop there, you’ll certainly get adequacy. But if you keep asking, you can get something far more insightful, nuanced and even original.
Here’s how that looks for me. I start with a pre–first draft: five or ten loose sentences to capture my initial thinking. No polish, no linear structure, just some loosely related ideas on page. Then I hand it to ChatGPT. The first output is usually smooth and organized, even publishable if originality isn’t the goal. (There’s a market for mere adequacy, btw, but that’s a different conversation.)
The real work begins in what comes next. I interrogate the draft. I might ask it to deepen a claim, connect two ideas it left separate, or shift the tone toward something more conversational. Or I might press for specifics: names, textures, examples. I challenge clichés and ask for counterarguments. Each question sends the exchange in a new direction and, crucially, sharpens my own thinking.
This becomes an iterative loop. At some point I’ll copy the evolving draft into a blank document, close ChatGPT, and work alone. That quiet step away is part of the interrogation too, where I integrate, reshape, and rediscover my voice. Then I bring the piece back for another round.
What emerges from this process is almost always stronger than what I’d produce solo. At minimum, the interrogation clarifies what the piece is really about. The earliest draft of this essay, for example, was expansive but diffuse. Sure, it was smoothly written and the insights were interesting, but it missed a conceptual core, what at Smarter Storytelling we call the Unifying Idea. By pushing on the AI’s responses, I could see what didn’t belong and concentrate on what did.
And that’s the real gift. Not speed, but insight. Interrogation forces you to decide what you mean. AI becomes a mirror that sharpens your ideas, sometimes by getting them wrong first and making you correct them.
So yes, AI can boost productivity. It can make us produce more in less time. But the deeper value is qualitative.
If you can treat it as an intelligent, if occasionally eccentric, creative partner, make the effort to interrogate it thoughtfully, your work can be more focused, more nuanced, and, paradoxically, more human.