LinkedIn's AI Slop Button Passes 1 Million Clicks

What happened

LinkedIn says more than a million users have clicked its "Seems like AI slop" feedback button since introducing it on 30 July. Chief product officer Hari Srinivasan confirmed the milestone in a post this week, describing the tool as a way for members to flag posts they suspect were generated or heavily assisted by AI with little added value. The button sits within the existing three-dot options menu on posts, alongside other reporting and feedback controls.

LinkedIn has framed the feature as a signal-gathering mechanism rather than a moderation or removal tool: clicks feed into the platform's ranking systems to help it understand which content members consider low-quality "slop," rather than automatically penalising or taking down flagged posts.

Why it matters

The story sits at the intersection of AI adoption and platform experience design. As generative tools make it trivial to mass-produce professional-sounding content, feed quality becomes a first-order product problem: platforms need mechanisms to separate genuine expertise and insight from formulaic AI output, without alienating users who legitimately use AI as a writing aid.

For any organisation running a content feed, marketplace or community, LinkedIn's move is an early data point on how platforms are operationalising trust at scale — turning member sentiment into a structured, ongoing input for ranking and content-quality models, rather than relying solely on post-hoc moderation.

By the numbers

• Over 1 million clicks recorded on the "Seems like AI slop" button since launch, according to LinkedIn's chief product officer.
• 30 July is the date LinkedIn introduced the feature.
• Three-dot menu placement means the control sits alongside LinkedIn's existing post-level reporting and feedback options, requiring no separate discovery step for users.

The Renascence take

The headline number is less interesting than what it implies about user behaviour: a million-plus clicks in a matter of weeks suggests people are actively, voluntarily policing feed quality when given a low-friction way to do so — a strong behavioral signal that platforms can design for.

Most coverage will treat this as an AI-detection story, but it's really a friction-reduction story: LinkedIn didn't build a smarter AI classifier, it built a one-click outlet for an emotion users already had. That's the behavioral-economics lesson — when frustration has nowhere to go, it curdles into disengagement; give it a visible, easy channel and it becomes usable product data instead. Experience leaders should ask not "how do we detect AI content" but "where are our own users already voting with friction, and have we built the button yet?" The risk LinkedIn now owns is expectation: a million clicks create an implicit promise that feedback changes what people see, and if the feed doesn't visibly improve, the button itself becomes the next thing members complain about.