AI made me a better writer (and slightly more annoying at dinner parties)

When AI writing tools showed up, I did what every confident writer does when confronted with a machine that can produce 1,200 words in twelve seconds.

I panicked.

Then I got curious.

Then I got slightly competitive.

Then I realised something.

A lot of the “AI writing habits” people complain about were… mine. All along. Sitting there in my Google Docs. Wearing a clever hat.

You know the ones.

• Repetitive sentence structures.
• Overly tidy transitions.
• Suspiciously balanced paragraphs that look like they’ve been meal-prepped.

Apparently that’s “AI-coded” now.

Which is fascinating, because I wrote like that almost two decades before AI had a login page.

So, let’s talk about it.

Because this whole AI moment hasn’t ruined writing. It’s forced us to look in the mirror. And that mirror is slightly unflattering but ultimately useful.

Pull up a chair.

The day “clear structure” became suspicious

For years, I taught writers to use clean structure. Logical flow. Short paragraphs. Clear transitions. Help the reader breathe. Don’t trap them in a Victorian sentence dungeon.

Now someone reads a well-structured article and goes:

“This feels AI.”

Oh. So clarity is robotic now?

I blame the overexposure effect. When you see something too often, you start noticing the pattern.

AI produces highly structured text at scale. Suddenly structure feels synthetic.

But humans love structure.

Read Malcolm Gladwell. Read Seth Godin. Read any competent columnist who understands rhythm.

They use repetition. They use deliberate phrasing. They build predictable patterns. It’s not laziness. It’s design.

AI didn’t invent structured writing, it just made it obvious.

And once it was obvious, we all had to ask ourselves: am I structuring for clarity, or am I structuring because it feels safe?

That question alone has made me sharper.

AI learned from us (yes, even the weird bits)

Before we start accusing the robots of inventing bad habits, we should probably remember something.

AI learned from humans.

It didn’t wake up one day and decide to overuse tidy transitions. It didn’t independently discover the joy of a well-balanced paragraph. It didn’t pioneer the dramatic rhetorical pivot.

It digested billions of pages of content. Articles. Blogs. Think pieces. Product descriptions. Academic papers. Forum rants at 2am.

The good. The bad. The aggressively mediocre.

And then it reflected patterns back at us.

So when we say, “AI always writes like this,” what we’re often describing is a statistical average of human writing.

Our habits, scaled. Our formulas, amplified. Our favourite sentence shapes, industrialised.

If AI has a quirk, it learned it from us.

Which is both reassuring and slightly embarrassing, because it means the mirror is accurate.

It also means we can’t blame the machine for everything.

If a certain phrasing feels tired, that’s not synthetic corruption. That’s collective human repetition finally made visible.

In a strange way, AI has done what good editors always do. It’s highlighted the patterns we stopped noticing.

And once you notice them, you can choose differently.

The repetition problem that wasn’t just AI

You’ve seen it online, “The problem isn’t X. The problem is Y.”

And suddenly everyone is side-eyeing that sentence shape like its contraband.

Fair.

But let’s not pretend humans didn’t love that construction long before ChatGPT arrived. It’s persuasive. It’s memorable. It feels decisive. It makes you sound like you know something.

Which is why I used it. A lot.

When AI began mass-producing those rhetorical shapes, the internet reacted. Now we’re hyper-aware of patterns and we can spot symmetry from 50 metres away.

And instead of sulking about it, I found it useful.

Now when I write something punchy, I ask:

“Am I saying this because it’s the best way to express the idea, or because it’s a familiar rhetorical trick?”

That pause matters.

AI didn’t create the habit. It made the habit visible.

Visibility breeds discernment and that’s growth, even if it stings a little.

The “too polished” anxiety

There’s a new cultural fear that if your writing is clean, confident and coherent, someone will accuse you of outsourcing your brain.

That’s funny, because for years I told writers to edit ruthlessly. Remove fluff. Tighten verbs. Kill adverbs. Channel your inner Strunk and White.

Now a smooth paragraph can raise suspicion.

We’re no longer comparing writing to average human effort. We’re comparing it to instant machine output.

Polish alone isn’t impressive. It’s expected. Which means your actual edge is voice.

Your lived experience. Your weird metaphors. The story about accidentally explaining narrative structure to a barista who just wanted to know your coffee order.

AI can imitate tone. It can’t replicate the specific flavour of your brain.

That’s comforting. Also slightly exposing because if your writing doesn’t contain you, then what exactly is it doing?

That question has made me more intentional.

I’m less interested in sounding competent. I’m more interested in sounding like me.

Which is occasionally less tidy, but far more human.

The myth of “AI tells”

People love to publish lists of “AI giveaways.” Certain transitions. Certain phrasing. Certain cadence patterns.

And yes, there are patterns.

But most of those patterns are just basic rhetorical habits. The kind you learn in school. The kind you unconsciously absorb from reading blog posts for a decade.

Humans cluster around effective structures. AI was trained on humans. It reflects those clusters back at us.

So when we accuse a sentence of being robotic, we’re often reacting to generic writing.

Generic writing predates AI.

AI just industrialised it.

Which, frankly, has been a gift. Because now we can’t hide behind “good enough” anymore.

If your piece sounds interchangeable, you’ll feel it. Your readers will feel it.

That discomfort forces evolution. And yes, I used that word deliberately.

How this changed my own process

Before AI, I’d draft, clean, polish, ship.

Now there’s an extra layer.

• I draft.
• I check for clarity.
• Then I scan for “default mode.”

Default mode is that slightly corporate, slightly symmetrical, slightly too-perfect rhythm that creeps in when I’m not paying attention.

Sometimes it’s fine. Sometimes it’s doing useful work. Sometimes that’s the brief.

But sometimes it’s camouflage. It’s me hiding behind technique instead of exposing insight. Playing safe when there really is no need to.

So now I deliberately inject asymmetry. A sentence that breaks pattern. A joke that undercuts polish. A moment of honest uncertainty.

Not for quirkiness, but for texture, because texture is hard to automate.

The humility factor

Here’s something we don’t talk about enough.

Writers got comfortable.

We built careers on being faster than most people at turning thought into prose. Then a machine arrived that can outpace us in raw output.

That can bruise the ego but it can also refine it.

Instead of competing on speed, you start competing on judgment. Taste. Selection. Perspective. Editorial instinct.

The real craft shifts up a gear.

The question isn’t “Can I write 1,000 words?” It’s “Should I? And which 600 actually matter?”

That’s a better question, and AI forced us to ask it.

The unexpected upside

The loud panic phase is fading. What’s left is a quieter recalibration.

• We’re more aware of filler.
• More aware of rhythm.
• More aware of voice as a differentiator.
• More aware that clarity and cliché are not the same thing.

Even the accusations help.

If someone thinks your writing sounds automated, you have an opportunity to examine it. Not defensively. Curiously.

• Where am I leaning on formula?
• Where am I playing safe?
• Where am I repeating because it works, not because it’s necessary?

Those questions improve craft and craft still matters.

A small confession

Sometimes I’ll run a paragraph through an AI tool and think: “Oh. That’s how I sound when I’m tired.”

It’s humbling but also clarifying.

The tool shows me a baseline version of my own habits. A slightly smoothed, slightly neutralised version of my style.

From there, I can decide.

• Do I want this sharper?
• More personal?
• More surprising?

AI becomes less of a rival and more of a diagnostic mirror. Writers who treat it that way are going to get better, not worse.

So what now?

If you’re a writer reading this, I’ll say something annoying but true.

This is good for us.

• It’s good that we can’t coast on structure alone.
• It’s good that polish isn’t rare.
• It’s good that formula is visible.
• It’s good that voice matters more than ever.

We’re being nudged toward intentionality and that’s where craft lives.

You don’t have to reject AI. You don’t have to worship it. You just have to become more conscious.

• More aware of your defaults.
• More deliberate with your rhythm.
• More honest about your patterns.

That awareness is uncomfortable at first, then it becomes power.

If that means I occasionally sound slightly less symmetrical and slightly more like the caffeinated human who actually wrote this, I’ll take it.

Because the real shift wasn’t machines learning to write. It was writers finally noticing how they do.

If you like the idea of working with a caffeinated human to improve your content, let’s talk!