LinkedIn's "AI Slop" Button Hits One Million Clicks in Two Weeks
Over one million LinkedIn users have clicked the platform's "Seems like AI slop" button within roughly two weeks of its launch, according to chief product officer Hari Srinivasan, who disclosed the figure in a follow-up post on August 20. The button, accessible from the three-dot menu on any post, was announced on July 30 as part of a broader package of measures to curb AI-generated content across the professional network. The Verge
Srinivasan had framed the effort as urgent from the outset. "AI slop is a top priority for all of us," he wrote on July 31, announcing that LinkedIn was ramping up the ability for members to flag posts or comments that appear to be machine-generated. He also acknowledged the definitional challenge at the heart of the effort, noting that "slop is hard to define." LinkedIn
The scale of the problem the button addresses is substantial. Pangram, an AI-detection firm, determined that 41 percent of LinkedIn's longform posts were flagged as fully AI-generated. That figure gives concrete shape to what had been, for many users, a vague sense that the platform's feed was filling with formulaic, machine-written content. The Verge
LinkedIn has paired the user-facing reporting button with backend changes. The platform introduced new classifiers — automated systems trained to identify posts as AI-generated — and removed a feature that would "enhance your post" with AI, effectively eliminating a built-in tool that contributed to the very problem it is now working to suppress. Srinivasan reported that users are "now experiencing 40% less views on what we classify as AI slop from just a few weeks ago," indicating that the classifiers are already feeding into distribution decisions, meaning the platform is showing flagged content to fewer people. The Verge
LinkedIn is also adding a feedback mechanism that tells users who make a flagged post: "Some members told us this post seems like AI." The message serves as a social signal to authors whose content has been identified by the community, creating a loop between reader reports and creator awareness. The Verge
This is not LinkedIn's first move against automated content. Earlier in 2025, the platform said it would crack down on comments users create at scale with little or no human involvement, targeting the kind of bot-driven engagement that inflates threads without adding substance. The Verge
The combination of measures amounts to a multi-layered approach: user-flagged reporting, automated classification, distribution penalties, a removed AI-authoring feature, and creator-facing feedback. Each layer addresses a different point in the content lifecycle, from creation through distribution to post-hoc identification.
The broader context here is that social platforms are now grappling with a content-quality problem that is partly of their own making. LinkedIn's decision to remove its own AI-enhancement feature is an implicit acknowledgment that built-in generative tools contributed to the degradation of feed quality. The 40 percent reduction in views for flagged content, reported within weeks of deployment, suggests that distribution throttling can move the needle quickly once classifiers are in place, though the durability of that effect is an open question. Classification accuracy is the load-bearing element; if Pangram's 41 percent figure is directionally correct, the volume of AI-generated longform content on LinkedIn is high enough that false positives — legitimate posts wrongly tagged as machine-written — could meaningfully affect real authors.
The one-million-click milestone, reached inside two weeks, signals that the user appetite for content-quality controls is real and immediate. Whether that engagement sustains over time, and whether LinkedIn's classifiers can keep pace with evolving generative models, will determine whether this is a durable shift or a corrective burst. The platform has at least committed the infrastructure to find out.
