The Hidden Cost of Letting AI Do Your Writing

AI is perhaps the most significant technology I’ve encountered in my lifetime (so far), and I use it for damn near everything: to get recommendations, wrestle with complex ideas, make better business decisions, and instantly solve technical roadblocks that used to take hours or days. I’m limited only by my ability to ask the right questions and my willingness to act.

But while it’s fun to get caught up in what AI can do, its widespread adoption carries costs.

I’m concerned about how the work landscape will change, and I feel for people whose careers will be made redundant. I also worry about AI’s impact on creativity and learning, especially through writing. Most people already avoid writing like the plague, treating it as a clunky communication tool that needs to be dabbled in but never mastered. Now that AI is here, writing feels one step closer to becoming a lost art.

Let me be clear: not all writing needs to come from your own brain. Here are some examples of writing tasks I’m perfectly happy to outsource to AI:

• Product descriptions

• Landing page copy

• Standard operating procedures (SOPs)

• Research summaries

• Idea generation

But for many other writing tasks, the act of writing itself is where the value lies. It’s how we clarify ideas, test our assumptions, and uncover what we actually think. Paul Graham puts it perfectly:

“If writing down your ideas always makes them more precise and more complete, then no one who hasn’t written about a topic has fully formed ideas about it. And someone who never writes has no fully formed ideas about anything nontrivial.”

Writing isn’t the only way to wrestle with complex ideas. An architect may sketch 3D models. A musician might improvise. A craftsman will use their hands. But writing is more accessible than any of those. And, to quote Graham again:

“However much you learn from exploring ideas in other ways, you’ll still learn new things from writing about them.”

Writing forces us to shape thoughts using the inherently limiting tool of language. The better you get at it, the better the alignment between what you know and what you can convey. Your clarity improves. So does your understanding.

I was reminded of this while reading Make It Stick by Roediger, Brown, and McDaniel. The book explores what makes learning effective, and two of its key principles — effort and elaboration — are exactly what writing demands. When you write, you retrieve, rearrange, and expand on ideas in your own words. That takes effort. When you revisit and revise what you wrote, you engage in reflection and spaced repetition, which deepens learning even more.

For people who only care about writing output — whether the final text communicates what they want — AI will be a godsend. But for those of us who write to understand, outsourcing that process is a mistake.

So yes, let AI write when it makes sense. But don’t let it think for you, too.