Here's something that should make you uncomfortable.

New research dropped this week, seven experiments, peer-reviewed, published in the Journal of Business Research, and the finding is brutal.

When people discover content was generated by AI, they don't feel mildly annoyed.

They feel moral disgust.

Not "oh, that's a bit generic." Not "I wish they'd tried harder."

Disgust. The same emotional category as being deceived.

Kirk and Givi (2025) ran this across seven separate experiments. The pattern held every time. AI-authored content triggers a perception of inauthenticity so strong that it produces a visceral emotional response, reduced loyalty, less word of mouth, and the kind of gut reaction people usually reserve for discovering someone lied to their face.

I read that paper three times. Not because I didn't understand it. Because I've been watching this happen in real time for two years, and nobody had the data to prove it until now.

Stay with me. Because this gets more interesting.

The researchers found the effect reversed under specific conditions. When AI was transparent, literally signing its own name, the disgust disappeared. When the content was factual rather than emotional, the reaction softened. When AI edited existing content rather than generating it from scratch, people didn't mind.

The backlash isn't against AI existing. It's against hidden authorship. Against the feeling that someone tried to pass off a machine's words as their own and hoped you wouldn't notice.

I've sat across from founders who've done exactly this. Posted AI-generated thought leadership for months. Got decent reach. Then wondered why nobody was converting. Why did the DMs dry up? Why did the content "perform" but nothing actually happened?

Now there's a word for what their audience was feeling. Disgust.

That distinction matters. A lot.

Think about how "organic" food labelling works.

It's not really about the food being objectively better in every measurable way. It's about the signal. Someone cared enough to do this differently. Someone didn't take the shortcut. The label communicates values before you've even tasted anything.

"100% Human" content is heading the same direction. A shorthand for: a person actually thought these thoughts. A person actually lived this story. A person actually sat with this idea long enough to say something specific about it.

CNN Business declared 2026 "the year of 100% human marketing." Merriam-Webster's Word of the Year in 2025 was "slop" — born from frustration with AI-generated garbage. In 2023, their word was "authenticity." The vocabulary is evolving faster than most marketing strategies.

The data keeps stacking. iHeartMedia's 2025 State of Audio report found 90% of listeners want human-created media. Pew Research Centre (2025) found 50% of Americans are now more concerned than excited about AI, up from 37% in 2021.

This isn't a niche opinion held by people who still use flip phones. It's a measurable, accelerating shift in how audiences evaluate content. And honestly? I felt it before the data confirmed it. You probably did too. That scroll-stopping moment where you read something and think this was written by no one. Not badly written. Just... hollow. Like biting into a strawberry that looks perfect and tastes like nothing.

Here's where it gets properly uncomfortable for anyone in my industry.

Haas, Kirk, and Givi (2025) ran five preregistered experiments and found that the discomfort isn't one-sided. Consumers who used AI to write emotionally personal messages and then presented them as their own felt guilty. The guilt was traced directly to perceived dishonesty.

They used Google's 2024 "Dear Sydney" ad as a case study. A father uses AI to write his daughter's letter to her Olympic idol. The backlash was immediate and savage. Not because the letter was bad, but because outsourcing emotional expression feels like deception, even to the person doing the outsourcing.

I think about that ad constantly. Because I've had prospects come to me after doing exactly this. Not with a letter to an athlete, with their own LinkedIn presence. Three months of AI-generated posts. Good engagement. And a growing, creeping feeling that they'd built an audience for someone who doesn't exist. The relief when they say, "I just want it to sound like me again", that's the guilt the research is measuring.

Brüns and Meißner (2024) found across three experimental studies that disclosure alone triggers negative reactions. Participants couldn't distinguish AI content from human content; the quality was identical, but the moment they were told it was AI-generated, evaluations dropped and perceived brand authenticity cratered.

The content was the same. The only thing that changed was the label.

Read that again. The quality didn't determine the reaction. The perceived origin did.

This is the thing that keeps me up at night. Not whether AI can write well. It can. But "well" isn't the point. The point is that content lives inside a relationship between the person who made it and the person reading it. AI doesn't break the quality. It breaks the relationship.

If you think this is all theoretical, France's Senate would like a word.

They're currently reviewing legislation targeting what lawmakers call "Pimping 2.0", their term for outsourced profile management and delegated user interactions on subscription-based platforms. The debate, covered by RTL, reflects a broader European push toward transparency requirements, intermediary accountability, and creator protection standards.

A Swiss startup called RedPeach conducted a six-month study analysing 2,000 publicly accessible creator accounts. Their finding: 65% of accounts generating $500-$100,000 in monthly revenue rely on external management structures or technological tools to manage user interactions.

Sixty-five percent.

RedPeach is now building facial verification technology for private messaging, designed to verify that responses originate from the account holder.

Their CEO, Marco Cally, put it plainly: "The creator economy has grown extremely fast. We believe the next stage is not only about growth, but about clarity. Technology should strengthen trust in interaction, not replace it."

Whether you think that's brilliant or dystopian, the direction is clear. We're heading toward a world where audiences want proof that the person behind the content is, in fact, behind it.

And I'll be honest, part of me finds this thrilling. Because I've spent my entire career in the gap between a person's real thinking and what ends up on the page. That gap is about to become the most scrutinised space in content. The people who know how to close it ethically are about to matter a lot more. The people who were papering over it with AI are about to have a very bad year.

This is the conversation my industry has been dodging.

If consumers experience moral disgust when they learn content was AI-generated, and if regulation is actively moving toward mandatory transparency, where does ghostwriting stand?

Here's my position. I'm not hedging it.

Ghostwriting is not deception. It's extraction.

Every speech a CEO delivers at a conference was shaped by a writer. Every book with a founder's name on the cover had editorial hands all over it. Every op-ed in the Financial Times went through a process. Ghostwriting has existed in every serious medium for decades. Nobody blinks.

The difference, the line that matters, is between extraction and fabrication.

Extraction: I sit with you for 90 minutes. I map how you actually think. Not the professional performance, but the way you explain things to someone you trust. The tangents. The analogies you reach for instinctively. The opinion you hold strongly but quietly. I build a voice profile from your actual words. Then I write content that sounds so unmistakably like you that people DM asking if you wrote it yourself.

That's the work. And it's hard. It requires me to disappear so completely into your way of thinking that my fingerprints aren't on the final piece. Not because I'm hiding, because the whole point is that it's yours.

Fabrication: Someone pastes your name onto AI-generated content you've never read, about topics you don't deeply understand, in a voice that isn't yours.

The first is a service. The second is the thing the research says triggers disgust.

The agencies running AI-first workflows, generating drafts with ChatGPT, lightly editing, and hoping the client doesn't notice, are on the wrong side of this. Windmill Growth's data from this month shows agencies using AI for first drafts have engagement rates 40–50% lower and client churn 2–3x higher than human-first writers. The market is already sorting this out.

I see it in my inbox. The number of prospects coming to me after leaving an AI-first agency has tripled since last year. They all say some version of the same thing: "It was efficient. It just wasn't me."

It's about what "human" actually means in your content.

"100% Human" doesn't mean anti-AI. I use AI every day for research, data analysis, scheduling, and pattern recognition.

What it means is: a human thought these thoughts. A human lived these stories. A human decided what to say and how to say it. The writing captures a real person's actual perspective.

Whether a human typed or a process assisted with the typing is secondary to whether the thinking is real.

Newman and Bloom (2012) found across five experiments that people value originals over duplicates through two mechanisms: performance (the uniqueness of the creative act) and contagion (the creator's essence transferring to the work). Granulo, Fuchs, and Puntoni (2021) found that this preference is strongest in symbolic contexts, when the content expresses beliefs and personality.

Contagion. I love that word for this. Because it's exactly what I'm trying to capture when I work with someone. Not their ideas in the abstract, their essence. The way they pause before the thing they actually mean. The phrase they always come back to. The conviction that sits underneath the professional polish. That's what transfers. That's what people feel when they read something and think this person is real.

Your LinkedIn content is symbolic. It's identity expression. It's how people decide whether to trust you before they've ever spoken to you.

The human essence in it isn't a nice-to-have. It's the entire value proposition.

Three things.

Audit your content for human fingerprints. Read your last 10 posts. Are there specific stories only you could tell? Numbers from your actual experience? Opinions you hold that most people in your industry don't? If every post could have been written by anyone in your field, that's not a voice. That's a void.

Decide where AI helps and where it replaces. AI for research, structure, and analysis? Brilliant. AI as the final voice? That's what the research says triggers disgust. Draw the line explicitly, for yourself and for whoever creates your content.

Get ahead of transparency. Regulation is coming. It may not hit LinkedIn ghostwriting directly for a while, but the cultural expectation is already here. If your process can survive scrutiny, if someone could watch how your content gets made and still trust it, you're fine. If your process relies on nobody looking too closely, that's not a sustainable position.

The "100% Human" label is coming. Like "organic," it'll be imperfect. It'll get co-opted. Some people will slap it on content that doesn't deserve it.

But the underlying shift is real.

People want proof that a human gave a shit.

The content that provides that proof.. genuinely, not performatively, is about to become the most valuable thing on the platform.

...or stay beige. Your call.

Written by Sarra Richmond, The Ghost.

I write the posts your favourite founders get credit for. Find me → linkedin.com/in/meetsarra