LinkedIn adds 'seems like AI slop' button as bot traffic eclipses humans

LinkedIn users can now flag posts they suspect were churned out by AI. The Microsoft-owned platform introduced a “seems like AI slop” button Thursday, a blunt new tool in what chief product officer Hari Srinivasan called a top priority for the company.

The move lands at a moment when the web is drowning in synthetic text. Cloudflare reports bot traffic has already surpassed human-generated requests — a milestone the infrastructure firm says arrived faster than it predicted. The frustration isn’t abstract. Digg’s Reddit competitor folded in March because it couldn’t control the bots. Substack just partnered with detection startup Pangram, which itself closed $9 million in new funding this week to fight AI-generated spam.

Srinivasan framed LinkedIn’s approach as multi-layered. Beyond the new reporting button, the company is deploying classifiers to demote AI slop in feed recommendations and using the flagging signal to train its detection models. It’s also blocking hundreds of thousands of automated comment attempts daily. Users whose posts get flagged will see a private dashboard notification nudging them to sound more human — a softer touch for people who use AI to polish rather than fabricate.

In a telling reversal, LinkedIn is scrapping its own “enhance your post” AI writing tool. That feature will be replaced by a simple proofreader. “People come to LinkedIn to connect with real people,” Srinivasan said. The subtext: even the platform admits its own AI tools were part of the problem.