One million people have pressed LinkedIn's AI slop button
Since late July, empty machine-written posts can be flagged. Uptake has far exceeded expectations — and reach for such posts has collapsed.
Illustration · AI-generated (AI IN LIFE)
At a glance
- Reporting feature introduced on 30 July 2026
- More than 1 million users have used it since
- Posts flagged as AI slop receive around 40 percent fewer views
- Pangram Labs: more than 40 percent of long-form LinkedIn posts are fully AI-generated — the highest of the platforms studied
- LinkedIn responds with prompts to authors rather than bans
LinkedIn introduced the feature on 30 July: in a post's options menu, alongside "not interested" and "report", users can now mark a post as likely AI slop — meaning text with little substance and little effort. According to the network, more than one million people have used it.
The effect is measurable. Posts in this category have received around 40 percent fewer views since the rollout. The chosen approach to authors is notable: LinkedIn does not ban and does not visibly penalise, but prompts users whose posts are frequently flagged about the pattern. The stance is corrective rather than punitive.
The backdrop comes from research by Pangram Labs: more than 40 percent of long-form posts on LinkedIn are entirely machine-generated — the highest share among the major platforms studied. That is no surprise on a network where regular publishing functions as career currency and text generators have driven the cost of it towards zero.
The real finding lies in what gets reported. Users are not flagging "written by an AI" but "says nothing". The two often coincide, yet they are not the same: a carefully researched piece written with model assistance rarely triggers the reaction, while three paragraphs of motivational filler with emoji bullets reliably does.
The button therefore captures a criterion automated detection handles poorly. Detectors ask where a text came from; readers ask what it is worth. For platforms currently betting on technical labelling, that is an uncomfortable signal: a million clicks do not measure how much AI is out there, they measure the audience's threshold for tolerating it.
FAQ
What exactly counts as AI slop?
Posts with low substance and low effort, typically machine-generated. What drives the report is emptiness, not proof that a model was involved.
Are reported accounts banned?
No. LinkedIn prompts authors rather than banning them. The visible consequence is mainly reduced reach.
Can I still use AI to write posts?
Yes. What gets flagged is the output, not the tool. Substantive writing produced with model assistance rarely attracts reports.
