A couple of years ago, I wrote a Six06 insight called "Is This the End of Writing?" It was pretty early on in the AI era and I was trying to articulate how I felt about AI as a tool for writers, and what I thought that tool might do to us, both now and in the long term.

Since then, much has been written about the opportunities and risks of AI-assisted writing, and it's become a polarizing topic. However, I don't think it's the black and white issue some people think it is. We can’t simply decide that AI writing is good or bad and then move on. There’s an incredible amount of nuance to it and part of that nuance comes from the type of writing that you’re producing.

Not All Writing Is Trying to Do the Same Thing

In grad school, I studied the essay. We talked about what an essay was and how it functioned, and we read what other writers had to say about it, too.

Famed memoirist Vivian Gornick once wrote that when she reads to find what an essay is about, she’s “looking for the inner context that makes a piece of writing larger than its immediate circumstance.”

That’s certainly true for the essay, but not every piece of writing is trying to convey a truth that's larger than its immediate circumstance. Some writing is simply trying to convey an idea, or a framework, or to provide context.

More than other forms of writing, there’s a practical utility in business writing that I think AI can help surface. On the one hand, if I publish a personal essay or memoir, I’m hoping that I can arrive at a universal human experience by bringing the reader into a specific one. Conversely, when I publish a business insight, I’m using a specific experience to surface a specific idea. There’s a difference.

And that difference shows that one rule for all of writing is inadequate.

Whose Writing Is It?

I also think there’s a distinction to be defined around what AI writing actually is. For example, if I have an idea for a piece of writing that’s based on my unique human experience, and I have ChatGPT interview me about that idea to crystallize my thoughts, and then after a robust Q&A the LLM generates a draft based on my thoughts, and then I revise that piece of writing, run it back through AI, and repeat that process a few times—well, whose writing is it? Is it AI’s? Is it mine? Is it slop that should be flagged by an AI detector? Where does my writing stop and AI’s writing begin? And is my idea any less valid than it would be had I written the whole thing unassisted?

Furthurmore, how then is that different from an executive hiring a ghostwriter who would essentially put him or her through that exact process? And what matters more, the process by which the writing came to be, or the idea the writing is trying to convey?

Which brings me back to that point about nuance—it's everywhere. If you type a prompt and hit Enter, it's probably not your writing. Everything else, however, feels like it's open to interpretation.

The Reader Still Comes First

I think what gets lost in most of the arguments around AI-assisted writing is this: the rules of writing still apply. If you intend your writing to be read by humans, then it has to be writing that's worthy of being read by humans. You have to, first and foremost, consider the reader. After all, that's who you're writing for.

I also think that if you, as the writer, are outsourcing the writing, then you become the de facto editor. There’s no scenario where you can outsource both of those responsibilities because if you do, you’ve abdicated the entire creative process and you can no longer claim any sort of authorship.

As the writer, you have an obligation to your reader. You have a duty. You also have an implied contract with them that says you will respect their time and deliver them an experience worthy of it.

So before you publish something written with AI, ask yourself one question: Have you honored your contract with the reader?