I finished writing the third and final draft of a piece and
out of curiosity (or mild self-sabotage) ran it through an AI detector. It came
back: 61% AI-generated.
Sixty-one.
Then, like a cheerful barista upselling oat milk, it asked: “Would
you like to humanize your text?”
I had to sit with that for a second.
Let me get this straight. I write the thing. The machine tells me it sounds like a machine wrote it. Then the same machine offers to make it more human. It’s like a plastic fruit offering to teach an apple how to grow on a tree.
I get that language has patterns. Maybe after years of
writing, my fluency itself is starting to look suspicious: too smooth, too
balanced, too … competent. As if being clear and structured is now evidence of
artificial origin.
Have we’ve reached a point where writing well can count
against you?
What bothers me isn’t that the tool got it wrong. Tools get
things wrong all the time.
What bothers me is the quiet confidence to essentially say: We
know what human sounds like now. And this isn’t it.
Since when?
But real writing isn’t a filter. It’s choices. It’s quirks,
blind spots, weird metaphors that somehow land. It’s the slightly crooked
sentence you leave alone because straightening it would kill it.
If anything, the more I’ve written over the years, the less
interested I’ve become in sounding “perfect.” Perfect is sterile. Perfect is
showroom lighting. Perfect is a couch no one is allowed to sit on.
So no, I didn’t click “humanize.”
_________________________
NOTE: For what it's worth, I posted this on LinkedIn and enjoyed over 1,300 responses,
over 225 comments, over 120 reposts, and over 50,000 impressions.