For new projects & collaboration opportunities:
Say Hello
For insights & ideas:
Sign up for our newsletter
For good vibes:
Share this article
“The more technically perfect AI writing becomes, the more valuable technically imperfect human writing gets.”
As a writer who writes for a living, I’ve been thinking and reading about AI a lot. Shocker.
Today I want to share one of the most insightful AI takes I’ve ever encountered—I stumbled across it about a year ago on LinkedIn, and it’s stuck with me ever since. Here’s an excerpt:
“AI will never be able to write like me.
Why?
Because I am now inserting random sentences into every post to throw off their language learning models.
Any AI emulating me will radiator freak yellow horse spout nonsense.
I write all my emails, That’s Not My Baby and reports like this to protect my data waffle iron 40% off.
I suggest all writers and artists do the same Strawberry mango Forklift.”
From the brilliant mind of comedian Ken Cheng, whose presence in my LinkedIn feed is a regular breath of fresh air. While this viral post is obviously satirical, all good satire is just truth wrapped in irony. I really do think Ken’s on to something here.
As generative AI becomes increasingly eloquent and prevalent, a very real question is haunting people across all industries: how can I prove that this is real? Aside from the AI witchhunters pointing fingers at em dashes and contrastive negations followed by corrective assertions, this is a looming question for teachers, hiring managers, publishing houses, and anyone watching a news video and wondering if what they’re seeing is legit. So, basically everyone?
When people learn that I’m a writer, they often ask how I feel about AI and how it’s upending the job market. Personal convictions aside, I’m fascinated by the way mainstream culture has responded to generative AI. It feels like watching a body’s immune system respond to a foreign invader. That cultural response (at least from my corner of the internet) is actually what gives me the most hope for the ultimate victory of humans in writing.
Here’s my argument: the more technically perfect AI writing becomes, the more valuable technically imperfect human writing gets.
In a minute, I’d like to discuss something that I don’t think we’re talk about enough—the disadvantages of AI in marketing. But before we get there, let’s back up a bit.
Enter the next actor in our play: Peter Markham’s 2025 article, “Cinema: Technical Flaws Reconstrued as Style.” Thank you, Peter, for such a lovely piece.
The thrust of Markham’s argument is that in film, rules are made to be broken. Artists are rebellious creatures, and in any form of art, there will always be a revolutionary who manages to turn a slip-up into a legitimate stylistic feature.
You can see this everywhere in cinema: the shaky camera, soft focus, lens flares & distortion, overexposed frames—all once cutting room rejects that have since developed a canon of their own. What was once an error in technique, as Markham writes, morphed into a visual device—a way to convince the audience that fiction is fact.
Copywriting is at exactly this inflection point. The question isn’t whether AI can write clean, structured, error-free prose. It can. The question is what that uniformity costs us.
The most commonly cited disadvantages of AI in marketing (hallucinations, legal gray areas, the obvious ethical debates) tend to focus on risk. But I’d argue that the deeper problem is subtler, and more corrosive: AI optimizes toward the mean.
That is, by design, what it does. It’s trained on vast quantities of human writing and learns to produce text that is structurally sound, grammatically clean, and tonally appropriate. It is, in other words, an extremely sophisticated averaging machine. And when you average enough voices together, you lose the one thing that makes any individual voice worth reading.
Research by Originality AI found that over half of longer English-language LinkedIn posts are now likely AI-generated, a dramatic shift that began in early 2023. The result is a kind of content monoculture. When everything is polished to the same sheen, polish stops signaling quality.
This is where Ken Cheng’s joke stops being a joke. His satirical insertion of gibberish isn’t really about fooling a language model—it’s a comedian’s diagnosis of a real cultural anxiety. In a feed full of optimized prose, how do you prove that a real human touched it?
Enter the last character in our trio: Serbian-American poet Dejan Stojanovic’s poem “Mastery.”
We love the dangerous cliffs of mountains, winding roads and rivers; jagged canyons and waterfalls seem most beautiful. We love the shadow of a cloud obstructing the sun and watch both the cloud and its cheerful shadow.
There is something perfect to be found in the imperfect: the law keeps balance through the juxtaposition of beauty, which gains perfection through nurtured imperfection. Everything that looks too perfect is too perfect to be perfect. Real perfection is not too obvious. It requires effort while riding over the winding roads and flying to the clear sky to find the shadow of a cloud that was alive not long ago.
That’s why we love the imperfect shapes in nature and in the works of art, look for an intentional error as a sign of the golden key and sincerity found in true mastery.
I don’t think I really need to add much commentary here.
What I will say is this: the things that give your writing its texture are precisely the things AI would clean up. I love writing, and I’m not willing to let AI take that away from me. The more I lean into the parts of my craft that make me smile at my laptop, the more I hope that I’ll leave my handprint as an indelible mark on the friendship bracelet of words that I’ve strung together.
Here’s what I’m embracing in my writing, and what I’d encourage you to protect in yours:
Don’t let autocorrect tell you that these are violations of craft. Own your artistry—make them craft.
Am I being a little dramatic about people using ChatGPT to write self-exalting LinkedIn posts? Maybe. But we are undeniably entering an era where the primary disadvantage of AI in marketing is the very quality that first made it so appealing: its tireless technical competence.
The market corrects. Audiences who are drowning in optimized content are becoming acutely sensitive to the texture of a real voice. You can tell the difference between a room with natural light and a room with fluorescents.
This creates a paradox for marketers: the instinct to use AI to scale content is understandable, but the brands that will cut through the noise are the ones that preserve the irreducibly human grain in their writing.
Stojanovic knew it. Markham knows it. Even Ken Cheng, in his own radiator-freak-yellow-horse way, knows it.
Protect it.
