LinkedIn is giving users a direct way to flag posts that appear to be low-effort, heavily automated or stripped of an identifiable human perspective. The professional network has added a “Seems like AI slop” option to its post-reporting menu, turning member feedback into a new signal for deciding which content should receive less distribution. The feature forms part of a wider effort to limit generic AI-written posts, automated comments and artificial engagement across the platform.
The move does not amount to a ban on AI-assisted writing. LinkedIn says it is trying to distinguish between people using AI to refine their own ideas and accounts publishing polished but repetitive material with little original knowledge behind it.
A New Feed-Level Control
The reporting option can be found through the three-dot menu attached to a post. After selecting it, users can indicate that the content appears to be AI slop. The reported post is then removed from that person’s feed, while the feedback is passed to LinkedIn as a signal that can help improve its content-ranking and detection systems.
LinkedIn Chief Product Officer Hari Srinivasan described the issue as a major product priority, saying: “People come to LinkedIn to connect with real people and share their real perspectives, ideas and expertise.”
The wording of the button is significant. LinkedIn is not asking members to prove that a post was generated by a specific model. Instead, users are being asked whether the content seems formulaic, automated or lacking in authentic input.
That distinction allows the company to collect feedback without presenting AI detection as perfectly accurate. It also acknowledges that poor content can be human-written, while useful posts may involve AI-assisted editing.
Human Reports Will Train Classifiers
Member reports will not operate as an isolated moderation queue. LinkedIn plans to use them alongside upgraded classifiers designed to identify generic AI-generated posts and other low-quality material.
When those systems identify likely slop, the content may receive less distribution through recommendations, especially when it comes from accounts outside a member’s existing network. The approach focuses on restricting algorithmic amplification rather than automatically deleting every post that uses AI.
LinkedIn disclosed in June that its early detection systems were correctly identifying generic content in 94% of its initial tests. The company said material that appears AI-generated and lacks a clear perspective is less likely to be widely distributed beyond the author’s immediate network.
The new reporting control adds human judgment to that process. A classifier can identify recurring sentence structures, vocabulary patterns and large-scale posting behavior, but it cannot reliably determine whether a post reflects genuine professional experience. Member feedback may help the system recognize the difference between assisted writing and empty content produced at volume.
AI-Written Posts Are Rising
The feature arrives as generative AI becomes increasingly visible across LinkedIn’s feed. An analysis published by AI-detection company Pangram found that more than 40% of long-form LinkedIn posts in its sample were classified as fully AI-generated. Separate figures associated with the research placed the share at about 41% for posts longer than 250 words and roughly 30% for shorter posts containing between 50 and 250 words.
These figures should be read as detector estimates rather than a definitive count of everything published on LinkedIn. Detection systems evaluate linguistic patterns and probabilities, and edited or mixed human-AI text remains difficult to classify with certainty.
Even so, the findings reflect a problem visible to many users: a growing volume of posts that repeat familiar leadership lessons, workplace observations and career advice without adding evidence, personal context or specialist knowledge.
On a professional platform, that creates a particular trust problem. LinkedIn posts can influence hiring decisions, business relationships and perceptions of expertise. If users cannot tell whether a claimed insight came from lived experience or an automated writing tool, the value of the feed begins to weaken.
LinkedIn Pulls Back Its Own Writing Tool
LinkedIn is also changing one of its own AI features. The company plans to remove its “enhance your post” tool, which could substantially rewrite a member’s draft. It will be replaced by a proofreading function intended to correct or refine the original language without replacing the author’s voice.
Srinivasan said LinkedIn had learned that many users run posts through AI because they want more confidence in their writing. The revised tool is designed to support grammar and clarity while leaving the underlying wording and perspective closer to what the member originally produced.
The change reveals the balance LinkedIn is trying to establish. It continues to treat AI as a useful writing aid, but it no longer wants its own product to encourage posts that sound detached from the person publishing them.
Creators May Receive Private Warnings
LinkedIn is preparing another feedback system aimed at people who publish content. The company will test private notices in creators’ analytics dashboards when members indicate that a post appeared inauthentic or relied too heavily on AI. Rather than publicly labeling the author or automatically penalizing the account, LinkedIn wants to show creators how their writing is being perceived.
This could be useful for people who wrote the central ideas themselves but used an AI tool so heavily that the final version lost their tone. It could also encourage brands and professional creators to review posts more carefully before publication.
However, the feature introduces a difficult moderation question. Perceptions of AI writing are subjective. A concise corporate style, formal English or repeated punctuation may lead users to suspect automation even when the content was written manually.
Members responding to LinkedIn’s announcement also raised concerns that competitors or groups of users could abuse the report option to target legitimate posts. LinkedIn has not publicly detailed how many reports would be required to affect distribution or how it will identify coordinated misuse.
Automation Beyond Written Posts
LinkedIn’s anti-slop work extends beyond post text. The company says it now catches hundreds of thousands of automated comment attempts each day and has blocked billions of other automation attempts over the past several months. These activities include mass posting, automated engagement and other behavior intended to manufacture visibility.
Automated comments are especially damaging because they can create the appearance of active discussion while adding little to it. LinkedIn has said it may exclude detected automated comments from its “Most Relevant” ranking and limit their visibility outside the commenter’s network. Accounts using prohibited automation tools may also face restrictions.
The company is pairing those defenses with expanded identity, workplace and company-page verification. More than 100 million members have added at least one verification to their profiles, giving LinkedIn another signal for distinguishing established users from fake or automated accounts.
A Shift From Detection to Quality
LinkedIn’s latest feature shows that the platform is moving beyond the narrow question of whether AI was used. The more important test is whether a post contains original context, credible experience or a useful point of view. A human can publish generic material, while an expert can use AI to improve a valuable draft. Treating all generated text as equally harmful would miss that distinction.
The reporting button therefore functions as both a feed control and a product experiment. LinkedIn will be testing whether member feedback can help its systems identify content that feels manufactured without unfairly suppressing legitimate writers.
Its success will depend on how carefully those signals are weighted. A well-designed system could reduce repetitive posts and restore space for real professional knowledge. A blunt one could turn subjective suspicion into an unreliable ranking factor. For LinkedIn, the challenge is no longer getting more people to publish. It is ensuring that the feed still contains something worth reading.
Comments