How to make AI sound less like AI

I didn’t set out to write this post.
I was just rereading something I’d published the night before, nodding along like a proud parent, then suddenly thinking, “Why does this sound like it was written by a very polite toaster?”
You know the feeling. It’s fine. It’s clear. It’s helpful. It’s also suspiciously smooth, emotionally flat, and just a bit too eager to please.
Like it’s wearing a name badge that says Hello, I am Content.
If you’re here, you’ve probably had the same moment.
You’ve used AI to draft something, maybe even something good, and yet it doesn’t quite feel alive.
It doesn’t breathe. It doesn’t sound like you. It sounds like… well, AI.
Let’s talk about how to fix that. Not with gimmicks or prompt yoga, but with actual writing decisions.
This is the behind-the-scenes version. The one where I admit what I mess up, what I rewrite, and what I quietly delete while pretending it was always intentional.
Pull up a chair.
First, accept an uncomfortable truth
AI isn’t bad at writing.
It’s bad at being human.
That’s not a flaw. It’s the point. It predicts language based on patterns.
Humans interrupt patterns. We ramble, backtrack, overexplain, undercut ourselves and occasionally say something slightly unhelpful but emotionally honest.
Most AI output fails not because it’s wrong, but because it’s too well behaved.
So, the goal isn’t to hide the AI. It’s to outgrow it.
You don’t make AI sound less like AI by asking it nicely. You do it by taking responsibility for the final voice.
That means editing. Actual editing. The unglamorous kind where you mutter “no one talks like that” and hit backspace.
Write like someone is listening, not grading
AI writes like its submitting coursework.
Clear structure. Balanced paragraphs. Neutral tone. No risks taken. It’s trying to get a good mark, not hold attention.
Humans write like someone might interrupt them.
Think about how you explain things to a friend. You don’t start with a thesis statement.
You start with context, or a story, or a slightly chaotic observation that makes sense two sentences later.
When I rewrite AI drafts, the first thing I do is mess them up a bit.
Shorten some sentences. Let one run on longer than it should. Add a thought in brackets that probably didn’t need to be there.
Not sloppily. Intentionally human.
If it reads like it’s been ironed, wrinkle it.
Stop explaining everything so perfectly
AI loves to fully explain itself. Every idea arrives with its coat neatly hung and its shoes lined up by the door.
Real writing doesn’t do that.
Sometimes you imply. Sometimes you trust the reader. Sometimes you leave a tiny gap and let their brain step in.
When something feels “AI-ish,” ask yourself this: am I over-clarifying?
Cut the obvious sentence. The one that restates the point. The one that begins with “This means that…” or “In other words…”
If the reader is smart enough to be here, they’re smart enough to connect the dots.
Let them.
Add opinions, even mild ones
AI avoids opinions unless you force it. It hedges. It balances. It presents all sides like a very calm mediator.
That’s useful for research. It’s deadly for voice.
You don’t need hot takes. You just need preferences.
Say you like something. Say something else annoys you. Say you changed your mind. Say you’re still figuring it out.
“I used to think X. Turns out I was wrong.”
That sentence alone does more human work than three perfectly neutral paragraphs.
Readers don’t connect with correctness. They connect with judgment.
Let the process show
AI presents finished thoughts. Humans reveal process.
One of the easiest ways to strip out the AI smell is to talk about how you got there.
The false starts. The thing you tried that didn’t work. The advice you ignored and later regretted ignoring.
This is especially true for AI content about AI. Meta honesty goes a long way.
I’ll often add lines like:
“I rewrote this section three times and I’m still not sure.”
That’s not weakness. That’s trust-building.
People don’t want perfect answers. They want to sit next to someone thinking out loud.
Break the rhythm on purpose
AI loves balance. Sentence lengths that politely alternate. Paragraphs that look good in a layout preview.
Humans don’t write like that when they care.
They stop short.
Or they go on a bit.
Sometimes they do both in the same paragraph.
When editing, read the piece out loud. Where do you sound like a brochure? Where do you sound like a person who’s slightly too invested in this topic?
Lean into the second one.
Reference things casually, not reverently
AI references ideas like it’s citing sources in a footnote. Very respectful. Very distant.
Humans reference things the way they reference songs or books they half-remember. Casually. Imperfectly. With opinions attached.
Instead of explaining a concept from scratch, you can say, “You know that feeling when…” and trust that they do.
It creates shared ground. AI struggles with that because it doesn’t share experiences. You do.
Use them.
Edit for friction, not smoothness
This one sounds backwards, but stay with me.
AI writing aims for frictionless reading. Nothing catches. Nothing surprises. You slide through it like a well-lit airport terminal.
Good writing has texture.
A sentence that makes you pause. A turn of phrase that feels slightly unexpected. A joke that lands a beat later than planned.
When editing AI drafts, I actively look for places to introduce tiny bits of friction. Not confusion. Texture.
That’s where voice lives.
The final pass is always human
Here’s the part many AI content creators skip.
They prompt. They tweak. They regenerate. They prompt again. Then they publish.
That’s how you get content that’s fine and forgettable.
The final pass should never be about improving the AI. It should be about reclaiming the piece.
Ask yourself:
- Would I say this?
- Would I argue this?
- Would I stand by this phrasing in a conversation?
If the answer is no, change it. Even if the original sentence was technically better.
Especially then.
A quiet closing thought
AI is an incredible drafting partner. It’s fast. It’s patient. It never gets tired of your nonsense.
But it doesn’t know why you care.
That part is still yours.
So don’t aim to make AI invisible. Aim to make your voice unavoidable.
If someone reads your work and thinks, “This sounds like a person who has opinions and probably overthinks their intros,” you’re doing it right.
And if you ever reread something and think, “Why does this sound like a toaster again?”
Congratulations. You noticed. That’s the whole game.
Now go wrinkle the page.



