Over a million uses in roughly three weeks. That number — quiet, specific, and a little damning — tells you more about LinkedIn’s feed problem than any internal report ever would. Since launching its “Seems like AI slop” feedback option on July 30, 2026, LinkedIn has watched users flag suspected Chatbots-generated content more than one million times. The platform is now claiming results: Chief Product Officer Hari Srinivasan says members are “experiencing 40% less views on what we classify as AI slop from just a few weeks ago.”
What the Button Actually Does
The mechanics are simpler than you’d expect — and more limited than you’d hope.
The feature lives in the three-dot menu on any post and sends a feedback signal to LinkedIn’s moderation systems. It doesn’t auto-remove content or flag the author as a spammer. Authors whose posts get reported receive a notification framing it as feedback, not a formal accusation. That framing matters more than it might seem.
Here’s what’s changed under the hood:
- The button is accessible via the three-dot menu on any LinkedIn post
- Flags function as classifier-training signals, not automatic takedowns
- Authors receive notifications describing the feedback as non-accusatory — LinkedIn’s framing is explicit: this is “feedback, not an accusation of spam”
- LinkedIn removed a feature that previously offered to AI-enhance posts before publishing
- AI-detection company Pangram found that 41% of LinkedIn longform posts it reviewed were fully AI-generated
The Limits of a Button
The data tells a more complicated story than a million clicks might suggest.
That 40% figure refers only to reduced distribution of content LinkedIn already classified as AI slop — not a documented drop in how much AI-generated content is actually being posted. LinkedIn hasn’t released any data showing overall AI content volume is falling. What you’re really looking at is this: users get the satisfying feeling of doing something; the platform gets better classifier-training data. Both things can be true simultaneously.
What LinkedIn still hasn’t disclosed is whether total AI content volume is actually falling — or whether creators are simply getting better at hiding the seams. It’s a bit like when YouTube removed the public dislike count in November 2021, keeping the button as a private signal while making the problem less visible. LinkedIn is at least feeding its signal back into its systems, which is more than YouTube did for transparency. Still, the company isn’t banning AI age laws-assisted writing outright. Creators using AI tools thoughtfully aren’t the target. The focus is distribution throttling, not authorship policing.
A million clicks is a real signal — one LinkedIn clearly needed. But if total AI content volume holds steady, the feed doesn’t automatically get cleaner. The platform may get sharper at spotting slop. Or creators adapt, and the seams just get harder to see. LinkedIn hasn’t answered which outcome is more likely.
