The Humanizers Strike Back
Reclaiming the Voice We Gave Away
I wrote The Humanizers in 2024 because something essential was slipping away in the rush to embrace AI for content creation.
Back then—in the explosive early days of widespread ChatGPT use—marketers, copywriters, coaches, solopreneurs, and creators were flooding the digital world with AI-generated text.
It was fast. It was cheap. It scaled effortlessly. But it felt empty. Generic. Soulless. Like a beautiful shell with no heartbeat.
Before AI came along I’d spent 20 years as a copywriter helping people connect through words that moved people to action, built trust, and created real relationships.
Then, like a thief in the night, AI came along and began producing content that could imitate structure and persuasion, but struggled to replicate the subtle empathy, the lived nuance, the idiosyncrasies that makes writing feel truly human.
So I wrote The Humanizers: Breathing Life Into AI-Generated Content as a practical guide and a quiet manifesto.
Through the SOUL framework—Sensitivity, Openness, Understanding, Leadership—I wanted to show how we could take raw AI output and infuse it with the emotional depth, personal stories, conversational flow, and authentic vulnerability that only a human could add.
It wasn’t about rejecting AI; it was about refusing to let efficiency erase the connection that only humans could create. The book became a bestseller because so many of us felt the same dread: we were gaining speed but losing soul.
That was 2024. Now here we are in March 2026, and the landscape has shifted in ways I didn’t see coming when I wrote that book.
So now, with the benefit of hindsight, I’m realizing that the crisis was never that AI sounds too human.
The crisis is that humans have begun sounding increasingly like machines.
And here’s the plot twist: we started doing this long before any LLM existed.
That’s right. We flattened ourselves first.
The Pre-AI Flattening We Started Ourselves
Go back to the mid-2010s. Long before “prompt engineering” was a job title.
SEO rewarded keyword-stuffed, scannable listicles with predictable headings and bullet points.
Twitter (now X) rewarded punchy, qualifier-free hot takes that maximized shares and minimized friction.
LinkedIn turned personal branding into polished corporate zen: short paragraphs, inspirational quotes, bullet-point wisdom.
Every major platform punished digressions, contradictions, long sentences, vulnerability without a clear CTA, or anything that didn’t convert or engage instantly.
So we adapted. We changed the way we write. And it wasn’t because we wanted to sound robotic. It was because the incentives were brutal: write legibly for algorithms, or watch your reach die.
We removed mess to keep the feed happy.
We standardized voice to stay visible.
We self-edited into frictionless, parsable prose so crawlers and search engines could love us more.
We were training ourselves to write like future training data.
Then the AI models arrived, scraped that decade of optimized sludge, mastered the patterns we’d mass-produced, and handed it back to us on a silver platter at infinite scale.
And just like that, forum posts, emails, even casual comments started reading like lightly prompted drafts.
Even Sam Altman noticed it. Studies tracked academic speech shifting toward the model’s favorite crutches—“delve,” “realm,” “intricate,” “underscore,” ad nauseum.
So we began consciously dumbing down our own writing just to sound “authentically human” again.
And then we turned around and used those same pre-flattened voices to “humanize” AI output.
It’s a perfect, elegant self-own, and we’ve been doing it on a global scale ever since.
This Is Why Humanization 1.0 Isn’t Landing Like It Used To
Look, the techniques from The Humanizers (those emotional arcs, personal anecdotes, conversational rhythm, the SOUL layers) still work when you apply them thoughtfully. I’ve seen them transform flat AI drafts into content that actually moves people.
But I’ve noticed something lately: even with all the right touches, a lot of pieces still feel a little distant. Like they’re missing that final spark of real human presence.
I think it’s because the human voice we’re bringing to the table—the one we’re layering on top—has already been quietly shaped by years of writing for platforms, not people. We’ve all done it to stay visible and relevant, which left some of our natural messiness and depth on the cutting-room floor.
So when we “humanize” AI output, we’re sometimes building on a foundation that’s already a step removed from the raw, unfiltered us.
Which brings us to the next level of humanization, which isn’t about adding more layers to the machine.
It’s about starting with a more reclaimed, less optimized version of ourselves first.
And the good news is, we can get that back.
This Is the Strike Back: A Recovery Operation
The Humanizers was Phase 1: breathe life back into the content.
This is Phase 2: breathe life back into ourselves.
We don’t need more polish on the output.
We have to make room for a rawer, less-sanded-down version of our voice before anything else.
Call it Humanization 2.0. Or just recovery.
The goal: break things. Produce writing so idiosyncratic, so full of friction, so un-parsable by agents and summarizers that it pushes back against AI’s easy remixing into the next synthetic loop.
A cognitive monkey wrench. Pockets of deliberate illegibility. Poison for the well of sameness.
Here’s what it looks like in practice:
Pre-Legibility Archaeology Dig up your oldest unfiltered writing—old blog posts from 2009, private journals, emails you never optimized. Notice the run-ons, unresolved thoughts, raw tangents, emotional whiplash without resolution. Ask: When did I stop allowing that? When did “clear and concise” become the only voice I trusted?
Anti-Optimization Rituals (Start These Weekly)
Write one full piece offline: no word count, no SEO ghosts, no audience in mind.
30-day ban on bullets. Force real paragraphs, real transitions.
Publish something deliberately “bad” by 2026 standards: 60-word opener sentence, zero buzzwords, one dangling question left hanging.
Monthly: Drop content too personal, too specific, too unmarketable for any crawler to summarize neatly. Watch what happens when you stop feeding the legible machine.
New Workflow: De-Optimize First, Then (Maybe) Involve AI Start with your raw, messy draft. Only then feed it to the model—and make it explain your unfiltered thoughts back to you. Force it to wrestle with real humanity instead of templating over it. Use AI as sparring partner, not ghostwriter.
This isn’t nostalgia for analog days.
It’s survival: refusing to keep speaking machine until we’ve remembered our own language.
The Movement That’s Already Writing Itself
If you’ve felt the hollowness creeping in—if your drafts STILL feel increasingly same-y, if you’ve caught yourself editing to sound “more human” (meaning more like the model)—you’re not broken. You’re awake.
My next book, The Humanizers Strike Back (coming in May), will expand this into the full toolkit: deeper archaeology, more rituals, and brand new strategies to poison synthetic data rather than feed it.
I’ll be giving you updates and wait list info as we get closer.
But the recovery doesn’t wait for the book.
Start today.
Because we never really “lost our voices to AI.”
We handed them over one optimized paragraph at a time.
And it’s time to take them back.



Interesting and I agree
Nice title! As with all tools and technology, AI is evolving and how we use it, think about it, etc., is too. Thanks for staying on top of where we need to go, Andy!