Overview
- LinkedIn confirmed on Thursday that it is testing a “Seems like AI slop” option in the three‑dot menu that lets users flag posts they suspect were written by AI and immediately hides those posts from the reporter’s feed.
- The company says flags will feed signals into new and improved classifiers that will downrank or limit AI‑style low‑quality posts in suggested content and help tune its models.
- LinkedIn is testing private dashboard nudges that tell posters their content “may have come off as inauthentic,” and it replaced its generative “enhance your post” feature with a proofreading tool that preserves a user’s voice.
- LinkedIn says it is already blocking very large volumes of automated activity—hundreds of thousands of automated comment attempts daily and millions of automation attempts in recent months—which it pairs with user reports to refine detection.
- Independent data from Pangram estimated roughly 41% of long posts and 30% of short posts on LinkedIn are likely AI‑generated, and commentators warn user flags could produce false positives, be misused, or become training data that changes detection and generative behavior.