LinkedIn's AI Slop Button Passes 1 Million Clicks

What happened

LinkedIn has confirmed that more than one million users have clicked its "Seems like AI slop" feedback button since the feature launched on 30 July. The button lets members flag posts they suspect were generated or heavily assisted by AI, giving LinkedIn a direct behavioural signal from its user base about the quality of content circulating on the platform.

Crucially, LinkedIn is not using the button to remove or delete flagged posts. Instead, the signal feeds into the platform's feed-ranking system, helping to determine how widely a given piece of content is distributed. The mechanism sits alongside existing moderation tools as a lighter-touch, crowdsourced input rather than a takedown trigger.

Why it matters

The move is a pragmatic response to a problem every major content platform now faces: distinguishing authentic, valuable posts from AI-generated filler at scale. Rather than building ever more sophisticated automated detection, LinkedIn is enlisting its own user base as a distributed sensor network, converting a subjective judgement — "this feels like slop" — into a scalable ranking signal.

For leaders in AI and digital transformation, this illustrates a broader shift: platforms increasingly treat user sentiment about AI content as a governance input, not just a satisfaction metric. It also shows how quickly a feedback mechanism can generate meaningful behavioural data once friction to use it is low — a single click, no explanation required.

By the numbers

• 1 million+ users have clicked the "Seems like AI slop" button since launch
• 30 July is the date the feature went live on LinkedIn

The Renascence take

The headline number is impressive, but the more interesting design choice is what LinkedIn didn't build: an automatic removal switch. That restraint reveals something about how platforms are learning to handle collective judgement on ambiguous content.

A single click with no removal power is a masterclass in low-friction feedback design — it captures a strong emotional signal ("this feels off") without asking users to make a definitive editorial call they're not qualified or willing to make. The behavioural insight here is that people will flag suspicion readily but rarely want to be the one who deletes something outright; separating detection from consequence lowers the bar for participation and yields far more usable data. Any organisation building feedback loops around AI-generated content, reviews or support interactions should study this pattern: let users signal discomfort cheaply, then decide centrally and transparently what that signal is allowed to change.