On 30 July 2026, LinkedIn added a new option under the three dots on every post: "Seems like AI slop." Tapping it hides the post for that reader and tells LinkedIn the content is not worth pushing to more people.
That button is not an AI detector. LinkedIn's own product lead said it plainly: AI and slop are not the same thing, and LinkedIn does not penalise you for using AI. Nothing decided which model wrote your paragraph. A person read your post, felt nothing, and tapped a button. So the thing worth fixing is not how your writing was made. It is the specific lines a reader bounces off, which is what this tool quotes back to you.
How to Use the LinkedIn AI Slop Checker
No account, no credit card, no upload. It takes about fifteen seconds.
- Step 01
Paste your post
Drop in the full post, not a summary. Line breaks matter to the score because shape is one of the five things being measured, so paste it exactly as you would publish it.
- Step 02
Read the score and the band
You get a total out of 20 and one of three verdicts: reads clean, some tells, or reads as slop. That bottom band is what reads as AI to somebody scrolling past. Every dimension gets its own score, so you can see whether the problem is what you said or how you said it.
- Step 03
Look at the quoted line
The report pulls an exact line out of your post and explains what a reader would react to. That quote is checked against your text before it is shown, so you never get handed a sentence you did not write.
- Step 04
Rewrite that line, not the post
Change the sentence that got flagged, then check again. Do not paste the whole post back into a generator: a second pass through the same kind of model pulls your writing further toward the average. Check which dimension scored low first, because it decides what will work. Specificity and surface tells live in individual sentences and a line edit fixes them. Shape, stance and reader payoff are properties of the whole post, and no line edit reaches those.
What This Checks That Other Tools Do Not
Most slop scanners count banned words. That was a useful signal in 2023. It stopped being one once every writing tool started stripping the same word list.
-
It scores substance, not just vocabulary
Specificity asks whether the post contains anything only you could have written: a number, a name, a date, a moment with a before and an after. A post can be completely free of banned words and still say nothing at all. That post is the one that gets reported.
-
It catches the shape, not just the words
One sentence per line. A short punchy opener followed by white space. Three of everything, then a tidy quotable closer. Readers recognise that silhouette before they read a word of it, because a shape arrives at scrolling speed while the words are still queuing up behind it. No word list reaches this, because it does not live in any single word.
-
It quotes real lines instead of giving you a percentage
Research on slop labelling found that professional editors barely agree with each other on a yes-or-no verdict, but they do agree on which specific spans are the problem. So this reports bands and quoted lines, never an invented "87% AI" number.
-
It never guesses whether a machine wrote it
AI-origin detectors misfire, and they misfire hardest on people writing in a second language. A Stanford study found 61% of TOEFL essays by non-native speakers were wrongly flagged as AI. This tool reports how writing reads. It has no opinion about who typed it.