LinkedIn added a button letting readers flag posts that look generated by artificial intelligence, reported at the end of July.
The decision says more about the state of content on the platform than about the technology. Marking something as machine-looking isn't an accusation of fraud: it's a complaint about text that adds nothing, and that tends to have a recognisable shape.
It's moderation by symptom rather than by origin. There's no reliable way to determine whether a text was generated, and the platform doesn't try: it asks readers whether it feels empty, which is a quality measure disguised as detection.
For anyone publishing with tool assistance, the practical effect is a new informal yardstick. The problem stops being using AI and becomes sounding like AI, which penalises repeated formulas, generic openings and conclusions that don't conclude.
The irony in the design is that the same audience complaining about generic text also produces it. The button measures collective fatigue with a format, and format is fixed by writing differently, not by hiding the tool.