AI has developed a thing all writers spend their entire careers striving for to perfect: a voice. An iconic, distinct way of speaking that is unique to it and could not be mistaken for anyone else.

This is not necessarily a good or even impressive thing.

Originally, I tried to write that intro as if it were in AI’s distinct voice. You know, a kind of jokey thing that would about make my point while also showcasing the voice.
After twenty minutes of this, I found myself on my apartment balcony, about to dig my eyes out with a pair of fountain pens and wondering exactly how I got here.

So instead, I’ll just quote Sam Kriss of the New York Times, whose piece “Why Does A.I. Write Like That?” inspired this one.

“In the quiet hum of our digital era, a new literary voice is sounding. You can find this signature style everywhere — from the pages of best-selling novels to the columns of local newspapers, and even the copy on takeout menus. And yet the author is not a human being, but a ghost — a whisper woven from the algorithm, a construct of code. A.I.-generated writing, once the distant echo of science-fiction daydreams, is now all around us — neatly packaged, fleetingly appreciated and endlessly recycled. It’s not just a flood — it’s a groundswell. Yet there’s something unsettling about this voice. Every sentence sings, yes, but honestly? It sings a little flat. It doesn’t open up the tapestry of human experience — it reads like it was written by a shut-in with Wi-Fi and a thesaurus. Not sensory, not real, just … there. And as A.I. writing becomes more ubiquitous, it only underscores the question — what does it mean for creativity, authenticity or simply being human when so many people prefer to delve into the bizarre prose of the machine?”

It says a lot that we can so easily identify this passage as intentionally mocking the A.I. voice. It says, in fact, that A.I. has developed its own voice!

It even has specific quirks and elements you can identify: a love of em-dashes and the trademark “It’s not X. It’s Y!” phrasing.
There are two reasons the development of a recognizable AI voice (across Claude, ChatGPT, Gemini) is bad for AI, but great for human writers.

Sidebar

For the record, as Sam notes later in the article, “It’s not X, it’s Y” is an extremely common rhetorical device. In fact, most of AI’s quirks are not unique to it, as in many ways it’s just averaging out all the writing it has vacuumed up over time.

But something about the way AI uses that rhetorical device makes me reach for the eye-gouging fountain pens again.

I can’t identify exactly what it is that’s so fountain-pen eye-gougey, and my wife has informed me I’m not allowed to study it further because, and I quote, “We only have so many spare eyeballs for you.”

But I think it’s that, like all its rhetorical devices, AI is not, and cannot, deploy them intentionally.

I See You

The first point of interest is that AI writing has become obvious. People are developing a sense of when an AI has written something, even if they can’t specifically identify the elements.

This detection isn’t perfect, and certain elements can trigger false positives — I am a long-standing lover of the em dash, and you will pry it from my cold, dead hands. But as an instinctive “spidey sense”? It’s good enough.

Unless someone is willing to put in some major sweat equity to edit it, AI-generated writing is apparent when we see it. Which, for a lot of lazy or bad actors, defeats the entire purpose of AI: its ability to firehose out whatever content you want or need.

In the Land of the Blind, the One-Eyed Man is King

Because the second point of interest is that for all its advancements (and I’ll admit it’s come a long way in what you can get it to do), AI writing still isn’t good. It’s barely even mediocre – it’s generic.
There is little worse than generic writing. I would actually posit that BAD writing is better than generic writing. Bad writing is, at the very least, memorable and often unique.

Ask any bad movie fan, and they’ll tell you that some of the greatest bad movie filmmakers have thumbprints as unique as the finest indie auteurs.

That’s crucial if you want to actually accomplish anything other than having Claude write mediocre Teenage Mutant Ninja Turtles fanfiction for you. Good copywriting, good storytelling, good marketing is about specificity.

You have a problem, a specific problem that’s causing you pain. My product, and no one else’s, is the solution to your problem.

Or as my copywriting mentor Phil Ruben put it: “The best copy format is: You have this problem. Wouldn’t it be great if there were a solution? I have that solution for you”.

If AI has an identifiable voice, if we can recognize that voice, that illusion is shattered: no person offering a solution to another person. Just a machine guessing the next word. Not speaking to you because it doesn’t know you exist. It doesn’t know anything.

It’s actually fitting that AI’s unique voice is the most generic one imaginable. Because the message being sent is one of vague blandness. It’s giving a message no one wants to hear in a style no one wants to read.

But you, out there in the digital wilds? You do know things. You do know people exist. You are specific.

And the land of the generic, even the partially specific person, is king.

A Word on the Truth

We live in dangerous times. You don’t need me to tell you that. You can look outside your window. You have the news. Armed, masked goons roam the streets, snatching whoever they don’t like the look of.

Predictably, they have killed in pursuit of this mission. In a few cases, “executed” is the only proper word for their actions.

But danger to life and limb is not the only danger lurking out there. There is also an epistemological danger. The concept of truth itself is under assault.

We have video proof of what I’ve just told you. A lot of video proof. And yet still some would tell a different story. A vile falsehood of a “terrorist” who kicked a car’s taillight and therefore deserved murder.

Not a majority, thank God, but still too many.

There are a lot of reasons for this assault on truth, but one of them is the cheap deployment of Generative AI, which allows propagandists and liars to push their filth out at high velocity. They don’t care about quality or accuracy, after all – they’re lying.

As I said recently to a fellow marketer, “You can sell anything if you’re willing to lie. Just not for very long.” Speed is key because going slow kills you.

So we must not give in to despair. Instead, we should take heart in two places.

1. This is possible because Generative AI is currently cheap for the consumer, but it’s almost impossible that it will stay that way forever. It costs too much on the back end, and the for-profit companies behind them will have to turn a profit or go out of business eventually.
2. AI has a voice. That means we can recognize it. That means we have another tool in our quiver to ask, “Are we being lied to?”

Stay safe out there.