The Real Reason People Want an AI Score on Your Writing
And it has nothing to do with transparency.
By now, a lot of you have seen the post from the Substack Team regarding their new AI detection tool. If you missed it, I attached a link below. Wander off to the comments, and you can obviously see people debating the use of this. Underneath it all, you start noticing a pattern — if you scroll through it long enough.
The pitch, on its face, is reasonable: readers deserve to know when they’re spending their attention on something no human actually thought about. Nobody wants to pour their evening into a stranger’s “reflections” only to learn later there was no stranger, or anything behind the words at all. That’s a real problem that deserves a real conversation.
But read the comments defending detection tools closely, and a different motive starts to show through the cracks. It has nothing to do with “I want to know what’s real,” and everything to do with “I want to decide what counts.”
This is the article:
Two very different complaints, wearing the same shirt
There are, roughly, two kinds of people cheering on AI-detection scores.
The first kind genuinely cares about being deceived. The sense of being fooled into investing feeling in something that had none behind it. That’s a coherent, defensible position. It’s about trust between a specific reader and a specific writer.
The second kind isn’t talking about trust at all. They’re talking about legitimacy. Their objection isn’t “I was deceived,” it’s “you didn’t earn this the way I did.” Scroll past the first few comments and you start seeing the tell: people who aren’t complaining about being fooled by anyone in particular, just insisting that a lower detection score should count for less, sell for less, and matter less regardless of what the writer actually did with the tool. Fraud complaint? More like a licensing scheme.
And you can hear the difference in the language, particularly in the second group.
The first group asks: was I misled?
The second group issues a verdict: this doesn’t qualify.
Purity tests don’t stay small
The gatekeeping crowd will tell you they just want transparency. But watch what happens next in these threads. The AI detection badge stops being information and starts being a hierarchy. Low score, high status. High score, guilty until proven innocent, and now you’re writing a defensive paragraph about your process just to be taken seriously.
One commenter on this thread put it well without meaning to make the opposing point: she’d published for a year and a half, used AI regularly, and not a single reader had ever accused her of being absent from her own work — because they could tell she was there, in the judgment, the rejection of bad lines, and the lived material only she had. Her point was that a detector can’t see any of that. It can only see word choice. If a movement treats word choice as the whole test of authorship, then it no longer protects readers from deception. The movement is now protecting a guild from competition.
This is what nobody wants to say out loud. If the tool actually measured “was a human present,” a lot of heavily-AI-assisted work would pass, because a lot of it has a person doing the parts that matter — like the arguments, editing, and the willingness to be wrong in public. What the tool actually measures is a text’s surface, and a surface is exactly what gatekeepers have always cared about: the diploma, byline, dues paid, and the years served. Now it’s a percentage instead of a pedigree, but the function is identical. Keep the door narrow. Keep deciding who gets to walk through it.
What actual authenticity-seekers would want
If the goal were really trust, this would look much different. It would be about disclosure, not disqualification — a writer saying plainly what role the tool played, and readers free to decide what that’s worth to them, the same way they already decide what a ghostwriter or an editor or a research assistant is worth. Nobody’s building a detector for how much your editor rewrote your sentences. Nobody’s slapping a badge on a book because a human research assistant did the digging. The outrage is oddly, and suspiciously, specific to one tool.
That’s the giveaway. This was never really about whether a person was there. It’s about which kind of person, working which kind of way, gets to call it writing.
And that, ladies and gentlemen, is the REAL reason people want an AI score.