"Seems Like AI Slop" Has Been Clicked a Million Times, and LinkedIn Says Views of Classified Slop Fell 40%

LinkedIn added a "Seems like AI slop" option to the menu on every post on July 30, for flagging content that looks AI-generated or low-effort automated. Chief product officer Hari Srinivasan said in late August that the button has been clicked more than a million times since launch and that, combined with classifier changes, views of what the platform classifies as AI slop have dropped 40% in recent weeks. Over the same period LinkedIn removed its "enhance your post" AI tool in favor of proofreading that doesn't alter your voice, and started privately telling authors when members think a post reads like AI. Worth noting: the 40% is platform-reported and measures reduced views of already-classified content, not less AI content on the platform — and how the classifiers decide has never been published.

A Million Clicks Measures Demand, Not Accuracy

The option sits in the three-dot menu on a post. Clicking it returns a thank-you and hides that post from your feed. Passing a million in a bit over three weeks says tolerance for template-voiced posts in the feed is low — there's no incentive attached, people are just annoyed. Report volume is not the same as correct judgment, and LinkedIn was explicit about two things: a single report doesn't decide a post's distribution, since the platform weighs "many signals," and it built safeguards so the feature can't be used to target someone unfairly. It also clarified that being reported doesn't necessarily reduce a post's reach. The 40% shift came primarily from updated classifiers, with reports as one input into tuning them.

How to Read That 40%

Srinivasan's phrasing is that members now see 40% fewer views on what LinkedIn classifies as AI slop than a few weeks ago. Look at the denominator: it counts **views of content already classified as slop**, not a 40% drop in AI-generated content on the platform. What the classifier gets wrong, and what it never catches, sit outside the number entirely. LinkedIn hasn't published its criteria, a false-positive rate, or an appeals path. There is one external reference point. Detection company Pangram scanned more than a million social posts through a browser extension between April 24 and June 30 and judged roughly 40% of LinkedIn longform posts to be fully AI-generated — the highest share among the platforms it measured, with Reddit around 4%. Detection has its own error rate, but together the two give a sense of the scale.

What This Means If You Write

Distribution is starting to discount anything that reads like AI wrote it, through a process that isn't transparent and has no clear appeal. That's the new variable for anyone publishing this year, and the practical response is unglamorous: don't open with the same construction every time, don't generate whole sections wholesale, and keep specific numbers, dates and first-person experience in the text. Those are both the fingerprints of human writing and the parts readers actually came for. LinkedIn may not be your main channel, but the mechanism is worth watching, because every large platform is working the same problem. "Platform-reported enforcement results plus an undisclosed classifier" looks like the shape this takes everywhere — and when it arrives, the only thing that vouches for your writing is the writing.

via: TechSpot, Engadget, TechCrunch on the button's launch