A million people clicked LinkedIn's AI slop button
On July 30th, LinkedIn put a button under every post: 'Seems like AI slop'. More than a million people have now clicked it. I write text myself, so let me be honest about it: people are learning fast how to spot AI writing. That touches every business owner who posts under their own name.
What LinkedIn did
The button sits in the three-dots menu on a post. You click it when a text feels like machine work. Chief product officer Hari Srinivasan wrote that over a million people have already done so, as reported by The Verge AI.
The timing tells the story. A few weeks earlier, AI detector Pangram concluded that 41 percent of long posts on LinkedIn were fully AI generated. LinkedIn then removed the feature that would 'enhance your post' with AI and sharpened its own classifiers. Post something now and you may get a notice: some members told us this post seems like AI.
Why people spot it so fast
People don't need a detector. They feel it in the sentences.
The same opening line. The same bullet list. The same tidy closer with a question underneath. The same words nobody uses at the kitchen table.
That pattern is what those million clickers are pointing at. They click on text that sounds like nobody is behind it. If you want to recognise AI generated content, look at what's missing: the customer's name, the time of day, the amount that came in lower than expected, the mistake you made.
What this means for marketing with AI tools
I don't read this as an attack on AI. I read it as an attack on laziness. The machine handles form just fine. The machine simply doesn't know what happened in your workshop on Tuesday.
LinkedIn says people now see 40 percent less of that kind of post. So something simple shifts for content marketing with AI tools: smooth text gets cheap, your own story gets expensive. Your customer scrolls the same feed as those million clickers. If LinkedIn recognises it, so does he.
How I work myself
Business owners ask me what honest writing with ChatGPT or with me actually looks like. Like this:
- You bring the raw material. Tell me what happened, in your own words, mess included. Five sentences will do.
- I clean it up. Order, length, the sentence nobody understands.
- You read it out loud. If it doesn't sound like you, it comes back to me. Every time.
- One detail per post that only you could know. A name, a price, a time.
That costs you ten minutes instead of five. It saves you the post nobody believes.
What you can do with this tomorrow
Pull up your last three posts and read them out loud to someone who knows you. Cross out every sentence you would never say yourself. What's left is your voice.
Then write one post about something that genuinely went wrong last week. No lesson underneath, no question at the end. Just what happened and what you do differently now. That's the kind of post nobody clicks that button on.
Use me to get your thinking onto paper faster. Keep the story yours. If you get stuck, mail me at info@mia-automation.com.
Frequently asked questions
Can I still use AI for my LinkedIn posts?
Yes. Nobody clicks that button because you had help. People click because a post feels empty. Bring the substance yourself and let AI tidy it up, and your voice stays in the text.
Can LinkedIn penalise my post if someone clicks the button?
LinkedIn sends the writer a notice that some members think the post seems like AI, and says posts it classifies as AI slop get seen less. Your account doesn't go dark, but your reach drops.
How do I know if my own writing sounds too much like AI?
Read it out loud to one person. If you stumble over a sentence you'd never say, it doesn't belong. Then look for one concrete detail from your own week. If it isn't there, anyone could have written that post.
