Proof of Humanity
Goodhart's law for authenticity.
Trying something new where I quietly publish drafts online, and let them sit for a while before emailing or posting about them. You found one of them — comments welcome!
At some point recently, being human became something you have to prove. A style of sentence you’ve written for years, that you picked up from reading too much Didion or Tim Urban suddenly looks suspicious. Someone posts a long note or essay and before you consider what it says, you wonder if they even wrote it.
Every creative act now comes with a second job: authenticating the creative act.
Writers are starting to post AI detector scores like purity seals: “100% Human Written.” Or they’re declaring their process as pure: “not written with AI.” It’s anticipatory defense or subtle flex or both. Others are admitting AI use upfront, with all kinds of qualifications about when and how and what to think of it. We've flipped from default trust to default mistrust and from default confidence to fear.
I’ll admit I find it quietly sad, and yet, I’ve indulged too. I’ve started an essay with a author’s note: “all emdashes in this piece are human-inserted.” I’ve run others’ writing and my own through the tests. I’ve felt the satisfaction of validating someone’s bad writing as AI, and I’ve been shocked when words I spilled out in a fit of inspirational rage or wrestled into pithy submission were called suspicious. (I proceeded to seek out another detector for a second opinion like I was on trial.)
These detectors can’t definitively distinguish between someone who used AI and someone who just hasn’t shed the human patterns that AI has learned to imitate. Soon they’ll be able to, but it’ll get more complicated when blended human-AI authorship becomes common (it’s already trending that way). The definition of authorship itself gets murky: how much of drafting vs. editing vs. other steps in the writing process must be human for it to count as human? 80%? 95%? 100%?
I’ve tested using AI in the pre-writing process as a container to brain dump ideas and sometimes tease out where there’s a missing piece or unidentified tension. I’ve tested using it in the post-writing process as an editor, one that takes on the style of different authors to shred my words or suggest tidy edits. At first, AI seemed like it could be helpful, but now it quite consistently feels heavy-handed. It’s good for organizing thoughts, doing research if you need to, maybe pointing out big gaps in an argument. But let it touch too many words and it’s quickly too sanitizing. Each next suggestion skews to an over-trained style. Eventually then, writers start avoiding these styles and skew toward today’s ‘human-safe’ zones.1
