Swipe the breakdown ›
For a year, LinkedIn put an Enhance with AI button on top of every post and all but begged you to press it.
Last week it added a different button. Seems like AI slop. Anyone can tap it, on any post or any comment, straight from the three-dot menu. It has been pressed over a million times in two weeks, flagged posts are getting 40 percent fewer views, and around the same time LinkedIn removed its own Enhance with AI button and hoped nobody would look too closely.
So the platform spent a year teaching people to write with a tool, then handed the crowd a way to punish the ones who listened.
Here is the part most people will get wrong. They will try to sound less like a robot. Out come the em dashes, off goes the word delve, in comes a detector, and the same empty post goes back up in a nervous new outfit.
But the button is not an AI detector. It cannot be. A stranger taps it in two seconds with no training, and what they mean is not a machine wrote this. It is this said nothing, and I could tell.
AI slop is not a technology problem. It is a nothing-to-say problem in a confident font. The tell is not the sentence structure, it is the genericness. If you could swap your name for a competitor and the post still stood, that is the slop, whoever typed it.
The fix is not to write more human. It is to write more you. I pulled every lead for one business recently, 415 of them. The 145 we had quoted closed at 62 percent. The 270 we never got round to quoting closed at 4.8. That is one sentence, it took real work to know it, and there is no version of it that belongs to anyone else.
Nobody flags that. There is nothing generic to point at.
The button is not really measuring how your post was made. It is measuring whether you had anything to say.