If you have scrolled LinkedIn lately and felt that creeping suspicion that a post was churned out by a large language model, you are no longer stuck just rolling your eyes. On July 31, 2026, the business-focused platform added a “seems like AI slop” button, letting members flag posts and comments they believe were cranked out by a machine.
The feature lives under the three-dot menu at the top of any post. Tap it, and you get a tidy confirmation — “Thanks for letting us know. Your feedback helps improve the feed” — after which the offending post disappears from view. It is a small interaction, but it turns passive frustration into actionable data.
LinkedIn’s Chief Product Officer Hari Srinivasan framed the move as a top priority. “We are ramping up a series of new and improved classifiers that identify if a post is AI-slop or generally low-quality content,” he wrote. “This will reduce the amount of AI slop you might see in suggested content and content from outside your network.”
The user-reporting angle is really about training. “Slop is hard to define and the definition changes,” Srinivasan added. “This lets us tune our models and make better feeds.” In other words, every tap teaches the machine what humans consider junk — a moving target that automated classifiers struggle to pin down on their own.
The context here is not subtle. LinkedIn has earned a reputation for hollow, formulaic posts that read as though an LLM wrote them — because many of them were. Some are pumped out automatically by bots, a headache that stretches far beyond one platform. Cloudflare, which provides a chunk of the internet’s infrastructure, says website traffic from bots now outnumbers traffic from actual humans.
Other companies are swinging back too. This month, Substack rolled out a tool that claims to flag when an email newsletter was written by AI. Detecting synthetic images and video is arguably further along, thanks to two systems:
- Google’s SynthID — an invisible watermark, imperceptible to people, that nudges pixels in a way Google’s tools can detect.
- C2PA — a cryptographically signed manifest that logs a piece of media’s provenance, recording details like the camera used, whether AI was involved, and what edits were made.
Written text is the tougher nut. There is no camera signature or embedded watermark to check, which is exactly why LinkedIn is leaning on crowdsourced human judgment rather than trusting classifiers alone. The button won’t magically purge every AI-authored humblebrag from your feed, but it does hand the community a lever it never had before — and gives LinkedIn a steady stream of real-world examples to sharpen its detection against.
Whether it meaningfully cleans up the feed or just becomes another button people mash in irritation, we’ll find out as the classifiers mature.
