LinkedIn's 'AI Slop' Button Has Been Clicked Over a Million Times
The platform says views on posts flagged as AI-generated have dropped 40 percent in just a few weeks, as users take the fight against filler content into their own hands.
Key points - LinkedIn users clicked the "Seems like AI slop" button over one million times since it launched on 30 July 2025. - Views on posts LinkedIn classifies as AI slop dropped 40 percent within weeks of the button going live. - Before the button launched, AI detector Pangram found that 41 percent of LinkedIn's long-form posts were fully AI-generated. - LinkedIn is now adding a private notice to post authors when enough members flag their content as AI.
LinkedIn added a pointed little button at the end of last month, tucked inside the three-dot menu on any post: "Seems like AI slop." The phrase describes low-effort, AI-generated content that clogs feeds without adding value, the digital equivalent of padding. As first reported by The Verge AI, chief product officer Hari Srinivasan confirmed Thursday that the button has been clicked more than a million times. We covered the launch on 30 July in "LinkedIn adds a 'seems like AI slop' button so users can flag fake-sounding posts".
Why did LinkedIn act now?
The numbers were hard to ignore. A few weeks before the button launched, Pangram, an AI-detection company, analysed LinkedIn's long-form posts and flagged 41 percent as fully AI-generated. That finding, first surfaced by 404 Media, put public pressure on the platform to act.
LinkedIn's response went beyond the button. The company rolled out what it calls "new and improved" classifiers, automated tools that scan posts and assign a probability of AI authorship. It also removed a built-in feature that rewrote a user's draft to sound more polished, one source of the very content people were complaining about.
What changes for ordinary LinkedIn users?
Two things matter here. First, posts the platform's classifiers label as AI slop now reach far fewer people, down 40 percent in Srinivasan's words. If you write something genuinely human and still feel like no one's reading it, the algorithm is at least less likely to be why.
Second, LinkedIn's rolling out a quiet nudge for authors. He framed it as helpful feedback, not a penalty, writing that the goal is to help people "not sound like AI" rather than to punish them.
Earlier in 2025, LinkedIn said it would crack down on comments generated automatically with little or no human input, so the platform's building several overlapping layers at once.
What should you watch for?
If you manage a company LinkedIn page and use AI tools to draft posts, this is a practical moment to review that workflow. A post that reads like a template is now more likely to be flagged by readers and buried by the algorithm.
For everyday users, the clearest sign the system's working will be a feed that feels slightly more human. Whether that happens depends on whether the classifiers keep pace with AI writing tools that are getting harder to detect.
The harder question, though, isn't whether a million clicks is impressive. It's whether suppressing flagged posts actually changes what people write, or just teaches them to write AI content that slips through.
Common questions
Can the AI slop button get someone's account penalised?
LinkedIn hasn't said the button triggers penalties. For now it feeds the classifier and, once enough people flag a post, sends the author a private notice.
Does this affect comments as well as posts?
Yes. LinkedIn separately said earlier in 2025 that it's targeting AI-generated comments created at scale with no meaningful human involvement, so both post feeds and comment threads are in scope.
