Since last week there is a "Seems like AI slop" button under every post. Blade Runner needed Deckard, a machine and a long list of questions to tell a human from a replicant. LinkedIn gives you one tap in a dropdown.
Here is the part I like. The button does not detect AI. It cannot. It says "seems like," then it hides the post and quietly trims its reach. So it was never an AI detector. It is a boredom detector wearing an AI costume.
That distinction matters to me, because I write a lot with AI. A lot. And I am not reaching for the delete button. Slop was never about who typed the words. It is about whether anyone thought before they were typed. A model can carry the sentences. It cannot carry the point. That part is still mine, and when I skip it, no prompt saves the post.
My favorite detail is the timing. In the same release LinkedIn retired its own "write with AI" assistant and replaced it with a proofreader. The platform that handed everyone the AI writing button now hands everyone a button to punish the results. Somewhere a product manager is having a very honest week.
A study making the rounds says 41% of long-form posts here are probably AI-written. The button will get a workout. I just hope people aim it at the empty posts, not at every post that happens to have a clean sentence in it.
So go ahead, run the test on this one. I wrote it with a machine. I also meant every word. I am curious whether the button can tell the difference, because that was always the only question worth asking.
LinkedWild