LinkedIn is testing a new button that lets users report posts as "seems like AI slop," according to the company's Chief Product Officer Hari Srinivasan. When someone flags a post this way, the person who wrote it gets a private notice through their LinkedIn dashboard suggesting that the post came across as fake or heavily AI-generated. Flagged posts will also show up to fewer people in their feeds (Engadget).

The reporting button will also help train LinkedIn's automated systems to recognize AI-written content (CNET). LinkedIn says it has recently improved those detection systems, so fewer AI-generated posts are being recommended to users. The company previously made an update to reduce how far AI-generated content spreads on the platform (Engadget).

At the same time, LinkedIn is removing its own AI writing tools. These tools had let users improve their posts and messages using AI that generates text. LinkedIn is replacing them with a simpler proofreading feature that checks spelling and word choice but does not change the user's writing style (Engadget). The removal of the AI writing assistance is documented in LinkedIn's help pages (LinkedIn Help).

The problem LinkedIn is tackling is large. A company called Pangram, which detects AI-written text, found that LinkedIn had more AI-generated content than any other social media platform it measured. More than 40% of longer posts on LinkedIn were flagged as fully written by AI (Engadget). In recent months, LinkedIn also caught hundreds of thousands of automated comments and blocked billions of attempts to post automatically on the platform (Engadget).

The new button follows reporting by 404 Media on how LinkedIn had become flooded with AI-generated content (404 Media).

Removing its own AI writing tools while introducing a button for users to report AI content is a notable reversal for LinkedIn. The platform had previously built AI writing help directly into its post composer, making it easier to create the kind of content it is now asking people to flag. Switching to a proofreading-only tool suggests LinkedIn concluded that its AI writing feature added to the authenticity problem rather than helping users.

There is a tension in how this system works that is worth thinking about. The "seems like AI slop" label is intentionally phrased as an opinion, not a confirmed fact. Someone flagging a post is not proving it was AI-written; the poster just gets a quiet notification. But that flag still has a real consequence, because the post reaches fewer people. So a post's visibility can be reduced based on other users' subjective impressions, and those impressions also feed into LinkedIn's automated detection systems as training data. The system is designed to get better over time, but early on it depends on the judgment of people flagging posts, and human ability to identify AI-written text is, at best, unreliable.

The numbers from Pangram help explain the urgency. If that 40% figure is accurate, someone scrolling through LinkedIn is seeing machine-written content at a rate that would have been hard to imagine two years ago. LinkedIn's enforcement figures, hundreds of thousands of automated comments caught and billions of automated posting attempts blocked, point to a platform dealing with large-scale bot activity and content farms, not just individuals using ChatGPT to help with their posts.

For people who use LinkedIn as their main professional networking tool, the practical effects are immediate. Posts that sound formulaic, overly polished, or templated now risk both social friction and fewer people seeing them, whether or not AI was actually used to write them. The reporting tool is still in testing, and LinkedIn has not said how many flags are needed to reduce a post's reach or whether there is any way to appeal if someone feels they were flagged incorrectly.