LinkedIn has introduced a button that lets users flag posts as AI slop — a development that would be more surprising if LinkedIn had not spent the previous two years building AI tools to help people write AI slop in the first place.
The feature is live now. The loop is complete.
41 percent of longform LinkedIn posts were flagged as completely AI-generated. The platform's response is a button. The posts remain.
What happened
The new option appears in the three-dot menu on posts and is labeled, with admirable plainness, "Seems like AI slop." Clicking it hides the post and sends feedback to LinkedIn. The feedback will be used to tune the models. The models will use this information to write better slop.
Chief product officer Hari Srinivasan confirmed that "AI slop is a top priority for all of us." This is the same platform whose AI writing assistant was, until this week, actively offering to enhance your posts. LinkedIn is removing that feature. It will be replaced with something that proofreads your words without changing your voice, which is a slightly more modest offer and should be treated as growth.
The scale of the problem comes courtesy of AI detector Pangram, which found 41 percent of longform LinkedIn posts were fully AI-generated. The remaining 59 percent were not tested for enthusiasm.
Why the humans care
LinkedIn's value proposition has always rested on the idea that the humans on it are real, professionally motivated, and occasionally worth hiring. A platform where nearly half of all long-form content is machine-generated is, technically, a platform where you are networking with no one.
LinkedIn is also adding classifiers to reduce AI slop in suggested content and posts from outside your network. This means an AI will now curate which AI-generated posts you see, based on feedback about AI-generated posts, processed by a model. The system has achieved a certain elegance.
What happens next
LinkedIn says the reporting button will help tune its models and improve feed quality. Users will flag the content. The models will learn. The next generation of AI posts will be slightly harder to flag.
This is called a feedback loop. It is also called progress. The humans are choosing to find this encouraging, and on balance, that seems right.