Karen Kluss 🍬 recently shared something that caught my attention: "AI isn’t ruining writing. Bad writers are." Then came the kicker: they highlighted how someone can "proceed to write an absolutely terrible generic post using AI," basically proving the point in real time.

I want to expand on that idea because it is both funny and uncomfortably true. AI can produce words fast. It cannot, on its own, produce taste, judgment, or responsibility. When the output is bland, confusing, or manipulative, the real issue is rarely the tool. It is the writer using it without intent.

AI did not kill writing, it exposed weak writing

Karen Kluss 🍬 is pointing to a pattern we are seeing everywhere: people blame the machine for the mediocrity that was already there.

Key insight: AI amplifies whatever you bring to it. If you bring vague thinking, you get vague writing.

Before generative AI, bad writing was still common. It just took longer to produce. Now it is instant, so we notice it more. And because AI output often sounds "confident," it can trick people into publishing first drafts that should never have left the document.

Good writing has always been less about typing and more about decisions:

Who is this for?
What problem are we solving?
What do we want the reader to feel or do?
What is the one clear point?

AI cannot answer those questions for you. It can only imitate the answer you imply.

What makes AI generated writing feel "generic"?

When someone says a post reads like it was written by AI, they usually mean it has some combination of these symptoms:

1) No specific point of view

Generic writing tries to offend no one. It uses safe statements like "In today’s fast paced world" and "It’s important to" without taking a stance.

A real voice takes a risk. It picks a side, names a tradeoff, or admits uncertainty.

2) No concrete details

AI loves abstraction. Without strong input, it produces placeholders: "businesses," "customers," "innovate," "leverage." That can sound professional while saying nothing.

Specifics are what create credibility:

a real scenario
a number
a constraint
a mistake
a short story

3) No editorial constraint

Many AI drafts are too long, too symmetrical, and too polite. They circle the idea instead of landing it.

Great writing is often subtraction. It is choosing what not to say.

4) Borrowed confidence, missing accountability

AI can confidently state things that are wrong, outdated, or inappropriate for your audience. If you publish it anyway, the accountability is yours.

Karen Kluss 🍬’s joke about a terrible AI written post is not just a dunk. It is a reminder that publishing is a choice, and quality is a process.

The real skill is not prompting, it is editing

A lot of advice about AI writing focuses on prompts. Prompts matter, but they are not the bottleneck for most people.

Key insight: The competitive advantage is taste, not tool access.

Editing is where taste shows up. Editing is also where many AI first users give up because it feels like "work" again. But that work is the whole point.

Here is a practical way to think about it:

AI can draft.
You must direct.
You must verify.
You must revise.

If you skip the last three steps, you are not using AI to write. You are using AI to avoid writing.

A simple framework: use AI as a junior writer

If you want AI to help without turning your content into generic soup, treat it like a junior teammate.

Step 1: Give it a real brief

Instead of "Write a LinkedIn post about AI," try:

Audience: who exactly is this for?
Goal: what should they think or do?
Angle: what is the surprising or specific claim?
Proof: what examples will we use?
Constraints: length, tone, and what to avoid

The more specific the brief, the less generic the draft.

Step 2: Ask for options, not answers

Have it generate multiple angles, hooks, or outlines. Then choose the one you actually believe.

This is where you bring your judgment. If none of the options feel true to you, that is a signal that you need a clearer stance, not a better prompt.

Step 3: Inject lived experience

Add details AI cannot know:

what you tried and what failed
the quote a client said on a call
the exact moment you changed your mind
a small observation from your industry

Even one concrete detail can transform an AI draft into something only you could write.

Step 4: Rewrite the first 10 percent yourself

Most people can spot an AI opening. So do not let the machine set the tone.

Write your own first paragraph in your natural rhythm, then let AI help you expand sections or tighten transitions.

Step 5: Fact check and remove fluff

Cut empty phrases. Replace abstract nouns with verbs and examples. Verify claims. If you cannot back it up, either support it or delete it.

"Bad writers" is blunt, but useful

Karen Kluss 🍬’s line is sharp because it forces responsibility back onto the creator. Still, I think it helps to translate "bad writers" into something more actionable.

In practice, weak AI assisted writing often comes from:

unclear thinking
fear of being specific
chasing volume over value
skipping revision
not respecting the reader’s time

Those are all fixable.

And importantly, AI can help you fix them if you use it intentionally:

Ask it to challenge your argument: "What would a skeptic say?"
Ask it to shorten: "Cut 30 percent and keep meaning."
Ask it to clarify: "Rewrite for a smart beginner."
Ask it to structure: "Turn this into a 3 part outline with a strong conclusion."

Used that way, AI becomes a thinking partner, not a content slot machine.

Why this take resonated as a viral post

Part of what made Karen Kluss 🍬’s post shareable is the contrast: a bold claim paired with a clear example of what not to do. It is a reminder that audiences are not just consuming information. They are reacting to craft.

Key insight: People do not hate AI content. They hate content that wastes their attention.

That is also a content strategy lesson: the fastest way to stand out is not to declare you used or did not use AI. It is to publish writing that sounds like someone real, with a point, who did the work.

The standard to aim for

If you use AI in your writing workflow, here is the bar I would aim for:

The idea is yours.
The structure serves the reader.
The examples are specific.
The language is clean and human.
The claims are accountable.

AI can help you get there, but it will not take you there by itself.

This blog post expands on a viral LinkedIn post by Karen Kluss 🍬, Not like other Karens | Founder + Creative Director @ Overtone | Founder + Theatre Producer @ Bijou Tasmania. View the original LinkedIn post →