LinkedIn Records One Million AI Slop Reports
LinkedIn recorded more than 1 million clicks on its "Seems like AI slop" reporting option within two weeks of its late-July launch, according to Chief Product Officer Hari Srinivasan. Srinivasan also reported that posts LinkedIn classifies as AI slop receive about 40% fewer views, while emphasizing that no individual report determines a post's distribution.
LinkedIn recorded more than 1 million clicks on its "Seems like AI slop" reporting option within two weeks of introducing it at the end of July, according to Chief Product Officer Hari Srinivasan. The option is available through the three-dot menu on posts for logged-in users.
Srinivasan reported that content LinkedIn classifies as AI slop is receiving about 40% fewer views than before the reporting option launched. He also stated that no single piece of user feedback determines how widely a post is distributed, because the platform uses a range of signals for ranking inside and outside a user's network.
What LinkedIn is measuring
Social Media Today reports that LinkedIn defines AI slop as content that may appear polished or sophisticated but lacks substance, experience, perspective, or insight. Under that definition, the reporting mechanism is not simply a detector for any AI-assisted writing. It is a feedback channel for posts users judge to be low-effort or low-value.
That distinction matters because generative AI usage and low-quality automated publishing are not technically identical problems. A post can use a writing assistant while still containing original domain knowledge, data, or analysis. Conversely, repetitive engagement bait and bulk-produced material can be low-value whether or not a model generated it.
The Register reports that Srinivasan described LinkedIn's existing automated enforcement as detecting hundreds of thousands of AI-slop comments each day, alongside billions of other automation attempts, including suspected mass posting, over recent months. The provided reporting does not give a methodology, error rates, or independent validation for those detection figures.
Feedback affects ranking, but not mechanically
LinkedIn's reported 40% view reduction should not be read as evidence that an individual "Seems like AI" report automatically suppresses a post. Both The Register and Social Media Today report that Srinivasan said distribution incorporates multiple signals rather than a single report.
According to Cybernews, LinkedIn is also adding analytics messages intended to show posters how viewers receive their content, including feedback indicating that users perceive excessive AI use. The available reporting does not specify the precise thresholds for those notices, the ranking weight of the feedback, or whether users can appeal an AI-slop classification.
For data scientists and ML engineers building content-quality systems, the rollout illustrates a familiar moderation-design problem: user reports can surface perceived quality failures at scale, but they are noisy labels. Platforms commonly need to combine reports with behavioral signals, anti-automation controls, classifier outputs, and human review to reduce abuse and avoid penalizing legitimate content.
A shift away from AI-assisted post generation
The Register reports that LinkedIn removed its AI-driven "enhance your post" feature in July and replaced it with a proofreader. That change occurred alongside the reporting option.
The result is notable less for proving that all AI-authored posts are harmful than for quantifying a large early response to a user-feedback tool. The 1 million-click figure measures reports or feedback submissions, not a verified count of AI-generated posts, unique offending accounts, or confirmed policy violations.
Across social platforms, comparable systems face a trade-off between suppressing scaled, low-substance publishing and preserving legitimate AI-assisted communication. LinkedIn's public statements establish that it is collecting this feedback and says distribution uses a range of signals; the provided reports leave open how accurately the system separates unhelpful generated content from useful professional writing.
Key Points
- 1LinkedIn logged more than one million AI-slop feedback submissions quickly, demonstrating substantial user engagement with a content-quality reporting mechanism.
- 2LinkedIn reported 40% fewer views for classified AI slop, while stating that individual reports do not independently control distribution.
- 3Platforms using user reports for AI-content quality typically need multiple signals because reports are noisy and cannot establish authorship alone.
Scoring Rationale
The rollout provides a notable real-world signal about user feedback, content ranking, and perceived low-quality generative AI material on a major professional platform. It is relevant to practitioners designing moderation and recommender systems, although the provided reports do not disclose detailed methodology or performance data.
Sources
Public references used for this report.
Practice with real Social Media data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Social Media problems
