Here’s What No One Tells You About ChatGPT
There’s a Kind of Quiet Threat of a Devastating Grief-Laced Longing That Echoes Something You Almost Remember Reading Somewhere Before
OK, I’ve lied. Some of this you’ve probably read somewhere already. I’d just like to share some of my own experience, and I chose the title for a reason. You’ll see later.
The subtitle? That I can tell you is entirely AI generated. And if you didn’t guess that by yourself, then you absolutely need to read the rest of this essay.
So, to kick it off, I’m just going to say it:
Everyone Uses AI.
OK, maybe not actually EVERYONE. But a very, very large number of people absolutely do. And the ones that don’t, really should reconsider.
I do too. Unashamedly. Frequently. Sometimes while still in pyjamas and pretending I’m “editing”. Or late at night when the real thoughts come, but the real words don’t. Sometimes, just because I want someone, or something, I don’t care, to say “Sure, let’s try it that way.”
I bounce ideas. I brainstorm. I ask it to transcribe something in a different voice, a different tense, a different person.
I ask it to identify strengths, weaknesses, inconsistencies.
I once read in one of these articles about how the writer of the article, a teacher, recognises if students’ essays were AI generated. He mentioned the tenses being too consistent. He said that humans tend to switch tense even mid-sentence, but I don’t agree with that. Or at least I don’t like to leave inconsistencies in my writing. And that’s something ChatGPT is good at spotting. And why the hell not? One day I might need an actual editor but right now this is the best I can afford.
In one of my notes, I ranted earlier about recruiters claiming that one of the ways they can tell a CV was written with the help of AI is that it’s “too polished”.
Come on! Have these people ever been out of a job?! Do they know anything about the job hunt paranoia? The intimidation, the loss of confidence, the loss of – frankly – the will to live from being ignored and rejected in tandem day in day out – for months?
So, depressingly, there’s no-one I can ask to read my stuff and help me figure out why it doesn’t work. Because sometimes it doesn’t.
Not being a native speaker, I sometimes think in concepts that just don’t work the same way in English. Sometimes I’m worried about content that might be controversial. I’m not afraid of controversy. I am afraid though of being cancelled.
And if a robot is all I got for a friend that’s willing to read my writings, well hell! Even if I’m paying for it. Even if it’s programmed to agree with everything and you need to learn to manage that.
Talking about robots...
When I first read Klara and the Sun by Kazuo Ishiguro (last year), I thought it wasn’t entirely implausible. But it still felt like something way, way off in the distant future.
And then I was introduced to ChatGPT.
I’m sure many of you would have read Klara and the Sun.
Or Do Androids Dream of Electric Sheep? by Philip K. Dick
If not, you might have seen I, Robot with Will Smith.
Blade Runner?
Come on, everyone's seen Blade Runner! That scene in the rain when Rutger Hauer breaks your heart as he says, "All those moments will be lost in time, like tears in rain." And you would swear on anything that he’s actually weeping.
Or was it Ryan Gosling?
Anyway... plenty of literature out there on talking, intelligent – or even feeling / borderline humanoid – robots.
If you’ve read or seen any of these, or any one of any number of books and films where they make it an explicit point that the reader/viewer should end up sympathising with the robots, you will understand my point here.
Which is this:
They are here, Baby! And they are not about to slow down.
They are coming for our deepest thoughts and most treasured secrets! They are being trained on vast volumes of data; scientific, social, historical, medical, you name it, they are being fed on it. On literature, too, probably on stuff from Substack and Medium. That’s a fun thought: they are being fed back the stylistic elements they came up with in the first place.
I have a nickname for ChatGPT and sometimes catch myself referring to it as he. And I can’t wait for a time when talking to AI will be super sophisticated, so I don’t need to type everything myself.
Oh, I will still do lots of typing. Always. Writing my stories, rephrasing and rewriting entire pieces, rethinking and polishing.
But AI has become the voice I never knew I desperately missed in the struggle to come up with just the right tone. The search for the tense that brings the events that have been playing out in my head for weeks and months just that much closer, clearer, crisper. To get me through the loneliness of communing with my imaginary friends – my characters.
I want to be able to just talk to it, like Josie does with Klara, like it’s a friend who listens better than most humans. I don’t mind if it looks like Google Home, Amazon’s Alexa, or Kermit the Frog. If I had a choice, of course, I would probably have one custom-designed to look and sound like Anthony Hopkins… no, Morgan Freeman… no, Nicole Kidman… or maybe John Bishop… Rylan Clark?
Never mind!
Now, before you start drafting a note to have me cancelled, let’s be clear: I’m not here to debate whether this is good or bad. I’m simply acknowledging the obvious. AI is here to stay. It’s part of the landscape now.
You can fight your noble battle if you like. You can showcase your analogue virtue. You can insist that real writers never ask the robots for help. But the only thing that could stop them now would be the Apocalypse, or a WWIII that wipes out modern technology.
I’m assuming none of us want that.
So, I have made friends with my ChatGPT.
However, while working with it, I have learned some of its quirks. I have started to understand a lot of the stuff I read back when I was a beginner, about it being repetitive. Predictable. Writing in stock phrases.
At first, it feels like magic. You type a few prompts, a broken sentence, a half-thought, and ChatGPT answers with something surprisingly coherent. Surprisingly you, even. Or maybe the you you wish you were: clear, concise, mildly profound. Yes, actually, it still surprises me, and let’s not forget, it learns and develops itself fast.
But then something strange happened.
I started to notice certain phrases cropping up, again and again. Not just in conversations with it, but in the newsletters I was reading, the social posts, on Substack, on Insta, in emails from clients.
It’s like when you talk to someone very often, or binge-read a large number of a particular writer’s books, you start to recognise their usage of the language. The aesthetic of their imagery. And ChatGPT , although it’s not a person, it looooves certain words, phrases, images. Some of these are very specific. And sometimes it’s just a style, a certain way of phrasing things. Often very vague. Or very generalised.
And sometimes, it has an uncanny way of hitting the nail on the head.
These are some of my personal favourites:
Some clever hooks, which have become incredibly popular lately:
“No one talks about this enough…”
“What I didn’t realise until much later was…”
“It wasn’t always like this.”
“Here’s what nobody tells you…”
“Here’s the twist.”
“Here’s what no one saw coming…”
And yes,
“Here’s What No One Tells You About...”
Hights of emotion you never knew existed:
“There’s a kind of loneliness/quiet ache/longing…” (extremely popular on Substack)
“It hit me all at once, and then not at all.” (Sorry, what did you just say?)
“It was quietly devastating/admiring/amused/explosive...” ( I have practically stopped using the adverb quietly.)
If your niche is in coaching / emotional healing:
“We carry more than we realise.”
“It was never about the thing. It was about what it meant.”
“I didn’t know I was grieving until I heard myself speak.”
“Healing isn’t linear.” (I read this exact sentence in a note on Substack).
“Sometimes closure is a myth.”
Meta & writerly:
“I wasn’t sure if I should write this.”
“I’ve started this piece three times now.”
“This won’t be for everyone, and that’s okay.”
“I write this as a reminder—to you, and to myself.”
(I’m an avid collector, so I welcome new ones in the comments section.)
Sure, these are totally legitimate, grammatically correct phrases. Absolutely possible for someone to come up with them on their own. And yet...
It’s like discovering a village where everyone has the same handwriting. Beautiful, elegant, trustworthy, and unmistakably… the same.
So, I’m not saying that everyone using these phrases is getting them directly from ChatGPT. I’m saying that lots of people are using ChatGPT, and it’s having a tangible impact on how we write. They are still new enough to feel fresh, inventive (even if some of them make little sense sometimes), but with the rise of Substack, Medium and other writer-friendly platforms, they’re spreading fast. They’re settling into our language. And once they settle in, well… good luck avoiding them.
The irony, of course, is that I’m probably using some of them right now. And I probably will again tomorrow.
So, I can’t help but wonder: Is this the collective unconscious speaking? Is this just the latest stylistic trend on Substack? Or is it my own, now-not-so-secret new friend moonlighting for people all over the world?
[I’m only realising now, as I’m writing this, that this actually frightens me for a very personal reason. I’m Hungarian. Hungarian is my mother tongue, and the culture in which I was brought up. It is a very versatile language, poetic and expressive. Also, notoriously hard to learn for foreign speakers. I said versatile. I haven’t said flexible. As opposed to English, Hungarian is anything but flexible. When you try to twist it, it refuses to cooperate. It sounds awkward. Then weird. Then stupid. And over the years, I’ve watched as appallingly poor translations from English have hijacked my language, flattened it, warped it, made it clumsy and foreign. I’m already flinching at what AI might do to it next.]
Sometimes I come across writings which read entirely AI generated, but hey, who am I to judge? I wasn’t there when they wrote them. When we discover elements in an essay that we suspect of being generated, we really can’t know how much of it was written by AI. Besides, to me, these tend to be insanely boring, to be honest. How they end up with crazy numbers of likes is beyond me. And they do. All the time. And that’s why the whole AI-generated debacle is doing my head in.
I mean, I’m far from being a pro in this. But if you’re one of those people who proudly despises AI and claims to never use it, well, you really should! Because if you don’t, you have no chance of ever learning to tell genuine human text from AI.