AI Tools
LinkedIn Just Admitted 40% of Your Feed Is AI Slop, Then Built a Button to Prove It
FlintPop Editorial · 4 August 2026 · 4 min
40%
Long-form posts flagged as AI-generated
0
AI rewrite tools LinkedIn kept, it killed its own
1
New button standing between your post and the feed
For two years, the entire pitch for AI in content was what you could make with it. On 30 July, LinkedIn spent its whole product roadmap answering a different question: whether people can tell.
The company's chief product officer confirmed what a recent study by AI-detection firm Pangram had already found, close to 40% of long-form posts on the platform are fully AI-generated. LinkedIn's response was not a policy update buried in a help article. It was a button.
"AI slop is a top priority for all of us," the executive wrote in the post announcing the change, the kind of line that would read as corporate boilerplate anywhere except underneath an actual product change.
The Button, and What It's Actually Attached To
A platform started grading whether your content sounds human. That is not a LinkedIn problem. That is every brand's problem now.
Every post on LinkedIn now carries a "Seems like AI slop" option in its three-dot menu. Click it, the post disappears from your feed, and the report feeds LinkedIn's own classifier models. That part is straightforward. The part with actual teeth is quieter: LinkedIn is testing a feature that privately tells you, inside your own analytics, when your audience believes your post was heavily AI-written or came across as inauthentic. For the first time, a platform is willing to tell a poster that their own numbers do not believe them.
Reach
New classifiers quietly reduce how far AI-slop content travels in suggested content and recommendations outside your network.
Trust
A private dashboard flag tells you, not the public, when your own audience reads a post as inauthentic or AI-written.
The swap
LinkedIn's own AI writing tool, "enhance your post," is gone. In its place: a proofreader that fixes grammar without rewriting your voice.
The Tell: LinkedIn Killed Its Own AI Tool
The detail worth sitting with is not the button, it's what LinkedIn removed alongside it. A platform does not retire its own feature unless it has concluded that feature was making the problem it's now fighting, not fixing it.
How it happened
Recent months
Pangram study: ~40% of LinkedIn long-form posts are AI-generated
30 Jul
"Seems like AI slop" report button ships to every post
30 Jul
"Enhance your post" AI writing tool retired
Rolling out
Private dashboard flag for AI-sounding content
Where This Leaves You
LinkedIn is the single most important organic channel for most B2B brands, and it just turned "does this sound like a person wrote it" into both a ranking signal and a button sitting in every reader's hand. That stopped being a writing-quality question. It's a reach question now.
Would your content survive that button?
If you're not sure whether your last ten posts would pass LinkedIn's own authenticity test, that's exactly the gap FlintPop's Brand Audit is built to catch. Free, ten minutes, no sales call.
Get Your Brand Audited →Want this applied to your brand?
Get Your Brand Audited →