LinkedIn is expanding its fight against AI slop by letting members report posts and comments that seem inauthentic and improving the automated systems that identify low-quality content. The company now faces a decision: how to reduce mass-produced posts and unwanted automation while treating legitimate AI-assisted work fairly.

Member reports will help LinkedIn refine the systems that recommend suggested content and posts from outside a person’s network. The company has not explained how much those reports will affect distribution or how creators can challenge mistakes.

The outcome affects members who rely on LinkedIn for professional knowledge and creators whose posts depend on the platform’s recommendations to reach new audiences. Independent creators and small teams may be especially vulnerable because research suggests people do not reliably recognize AI-generated writing. Readers can interpret AI-assisted polish as evidence that a creator’s ideas or expertise are inauthentic, even when the work reflects genuine knowledge and human judgment.

In short: LinkedIn wants to protect the professional knowledge that makes its platform useful. Its challenge is to separate low-value automation from genuine work created with AI assistance without allowing suspicion alone to determine what gets recommended.

AI slop generally refers to low-value, mass-produced AI content, but LinkedIn says the category is difficult to define and has not disclosed how it will separate slop from legitimate AI assistance.

An AI slop reporting system combines member reports with automated classifiers to identify low-quality content and help a platform improve what it recommends.

People come to LinkedIn to connect with real people and learn from their perspectives, ideas, and expertise. Hari Srinivasan, LinkedIn’s chief product officer, said protecting that experience from AI slop is now a top company priority. Mass-produced posts and automated comments can crowd out the professional knowledge people expect to find, making the feed less useful and harder to trust.

LinkedIn says it catches hundreds of thousands of automated comment attempts every day. Over the past few months, the company has also blocked billions of other automation attempts, including accounts posting at scale and spreading slop.

The company is now expanding members’ ability to flag a post or comment that seems like AI slop. That puts members in the position of judging a category LinkedIn says is difficult to define and whose meaning changes. Srinivasan stressed that “AI and slop are not the same thing” because many people use AI to refine their own thoughts, but LinkedIn did not provide criteria for separating legitimate assistance from slop.

If LinkedIn must remove low-value automation without treating every use of AI as suspicious, how will a button based on what content “seems like” slop keep those categories separate?

LinkedIn is also improving the classifiers it uses to identify AI slop and other low-quality posts. These automated systems will focus on suggested content and posts from outside a member’s network, reducing how often LinkedIn recommends material it considers low quality. Member reports will provide another source of feedback that LinkedIn can use to tune those systems and improve its feed recommendations.

The company has not said that reporting an individual post will automatically reduce its visibility. It also has not explained how much weight reports will carry, whether content will be reviewed before its distribution changes, or whether a report could affect the creator’s future posts. Any reduction may apply only to recommendations outside a member’s network, but LinkedIn has not clarified whether followers could also see less of the content.

LinkedIn also plans to test private notices in creators’ analytics dashboards when members perceive their posts as inauthentic or heavily AI-assisted. Those notices would reflect readers’ impressions rather than a confirmed finding about how the content was produced. Both the reporting feature and the updated classifiers are being expanded, while the dashboard notice remains a test, so LinkedIn has not established how widely these measures will be available.

If reports can help shape what LinkedIn recommends, even when their effect on an individual post remains unclear, how accurately can members recognize AI-generated writing?

Research into 25 million comments posted on Hacker News and Reddit between 2023 and 2026 suggests that people are poor judges of whether writing came from AI. The researchers used an AI model to assess 7,500 sampled accusations, manually categorized 300 instances in which users accused others of using AI, and compared accused comments with similar comments that had not drawn accusations.

Pejorative accusations of AI use increased more than tenfold on both platforms during the period studied. “AI slop” accounted for 94% of those negative labels, as accusations increasingly became a way to challenge whether someone’s contribution was authentic.

What purpose does an accusation serve if it does not reliably identify AI writing? The study found that people were not reliably recognizing the writing patterns associated with AI. Among human-written comments, the patterns that statistically distinguish AI text from human text did not predict which comments would draw accusations. The study does not establish exactly why individual users made those accusations, but the researchers concluded that people were often reacting to whether a comment felt authentic or belonged in the community. An accusation can then become a form of community enforcement: deciding whose contribution should be treated as legitimate, protesting the amount of artificial or low-effort content on a platform, and showing that the accuser understands and defends the community’s standards. The researchers describe this as social gatekeeping. A detection tool may assess how something was written, but it cannot settle that social judgment.

The research examined conversations on Hacker News and Reddit, so it does not show how LinkedIn’s reporting and classification system will operate. It also does not establish that LinkedIn will misclassify a particular post. It does show why a member’s suspicion alone should not be treated as proof that AI produced a post.

If a mistaken accusation can enter a system that helps decide what gets recommended, how will legitimate creators learn what happened and challenge an incorrect judgment?

LinkedIn says it wants creators to receive authenticity feedback from real people because an AI detector can make mistakes. Member reports will help the company tune its models, but LinkedIn has not said that a single report will automatically suppress the post it targets or explained how much influence an individual report will have. The only notice described in the announcement is the private dashboard flag LinkedIn plans to test when members perceive a creator’s work as inauthentic or heavily AI-assisted.

That leaves creators with little information about what happens after a report. LinkedIn has not said whether it will tell them when reports or classifiers reduce their reach, show them the reason or evidence behind a warning, or reveal how many reports their content received. The announcement also does not explain how creators could challenge an incorrect decision, how LinkedIn will prevent coordinated or malicious reporting, or how mistaken signals would be removed from its systems and future model training.

Substack has taken a more explicit approach to AI detection on its platform. Through a partnership with Pangram, readers can request an estimate of how much of an eligible post, note, reply, or comment appears to be human-written or AI-assisted. The result is shown only to the reader who requests it, and Substack has not said that the score affects the content’s visibility or recommendations.

Creators can add a “How I make this” statement explaining their writing process, which appears when someone scans their work. This lets readers see how the writer used AI before assuming that a high AI score means the post lacks genuine human thought or effort.

A creator who believes Pangram produced an inaccurate result can report the detection error to Substack and explain why the analysis is wrong. This challenges the detector’s result; it does not report the creator’s post for moderation.

Separately, creators can disable detection on one of their posts or notes. Readers will no longer see a Pangram analysis, but they will be told that AI detection is unavailable. This gives creators control over whether an analysis appears alongside their work, although disabling it could create suspicion because readers are not told why the creator made that choice.

Those safeguards do not establish that Substack’s detector is accurate, but they give writers ways to explain their work and challenge a result. LinkedIn has not described comparable options for creators affected by its member reports, private warnings, or classifiers.

Without clear standards or a way to challenge mistakes, the greatest risk may fall on independent creators and small teams that use AI to turn genuine expertise into professional content.

LinkedIn acknowledges that some members run their writing through AI because they feel more confident publishing the fuller posts expected on its platform. The company also recognizes that AI can refine a person’s own thoughts without replacing the ideas behind them.

For a freelancer, startup, or small business, AI assistance can provide writing, editing, or design help they may not have in-house or be able to hire. Larger organizations can pay employees or agencies to provide comparable support. If AI-assisted polish becomes a negative authenticity signal, smaller creators could face suspicion for using a tool that helps them compete with better-resourced organizations.

The risk becomes more serious when readers cannot reliably identify AI writing. As the research in the previous section showed, an accusation can reflect whether someone’s work feels authentic or belongs in a community. A creator could therefore be judged for how a post sounds even when the ideas, expertise, and final decisions are their own.

An unclear standard and no clear way to correct mistakes could put creators with fewer resources at a disadvantage, although that effect has not been measured.

A fair standard cannot simply ask whether AI helped produce the words. It must consider whether the post contributes genuine knowledge and whether its creator can challenge an incorrect judgment.

LinkedIn has offered an “enhance your post” feature that uses AI to help members write posts and messages. The company is now removing that feature and says it will introduce a tool that proofreads a person’s words without changing their voice. LinkedIn has not described what the new tool will correct or said when the change will happen.

The decision grew from LinkedIn’s effort to understand why members use AI to write. Srinivasan said some people feel more confident running their words through AI because LinkedIn expects more than a short response. The replacement feature appears intended to help with the writing while preserving more of the member’s original language, although LinkedIn has provided few details about how it will work.

LinkedIn is also expanding verification for profiles and company pages so members have more information about who they are connecting with and who is behind the content they read. Profile verification is optional, while eligible page administrators can request company-page verification for LinkedIn to review. These badges can confirm information such as a person’s identity or workplace, or that a page officially represents an organization, but they do not verify who wrote an individual post. LinkedIn has not provided a timeline for the expansion.

Members will also be able to block comments from company pages they no longer want to see. Srinivasan said LinkedIn has more work planned to protect its feed from slop, unwanted automation, and content that does not reflect genuine professional perspectives.

LinkedIn will need clear standards, transparency about whether and how reports affect distribution, and a way for creators to challenge mistakes. The real test is whether it can reward genuine contributions and stop harmful automation while treating AI-assisted work fairly when the ideas and judgment come from a real person.

Q: What is LinkedIn doing to reduce AI slop?
A: LinkedIn is expanding members’ ability to report posts and comments that seem like AI slop. It is also improving the automated classifiers that identify low-quality content and testing private dashboard notices when members perceive a creator’s work as inauthentic or heavily AI-assisted.

Q: Will reporting a LinkedIn post as AI slop reduce its reach?
A: LinkedIn has not said that one report will automatically reduce a post’s visibility. Member reports will provide feedback that LinkedIn can use to refine recommendations for suggested and out-of-network content, but the company has not explained how much weight reports will carry, whether posts will be reviewed before their distribution changes, or whether followers could also see less of the content.

Q: Does LinkedIn consider all AI-assisted writing to be AI slop?
A: No. LinkedIn says AI and slop are not the same because people can use AI to refine their own thoughts. The company has not disclosed the criteria it will use to separate legitimate AI assistance from mass-produced or low-value content.

Q: Why is LinkedIn changing its own AI writing tools?
A: LinkedIn says some members use AI because they want help producing the fuller posts expected on the platform. The company is removing its “enhance your post” feature and plans to replace it with a proofreading tool intended to preserve the member’s voice, although it has not explained what the new tool will correct or when it will become available.

Q: Can people reliably recognize AI-generated writing?
A: Research examining comments on Hacker News and Reddit found that people did not reliably identify the writing patterns associated with AI. Accusations often reflected whether a contribution felt authentic or belonged in the community, although the research does not establish how LinkedIn members will use the reporting feature.

Q: How could LinkedIn’s AI-slop system affect small creators and businesses?
A: Independent creators and small teams may use AI for writing, editing, or design support that larger organizations can obtain from employees or agencies. If AI-assisted polish becomes an authenticity warning, smaller creators could face suspicion even when the expertise, ideas, and final decisions are their own.

Q: What can creators do if LinkedIn wrongly flags their work as AI slop?
A: LinkedIn has not explained how creators could challenge incorrect reports, classifier decisions, or private authenticity warnings. Substack allows writers to explain how they created their work, report an inaccurate AI-detection result, and disable detection on individual posts, but LinkedIn has not described comparable safeguards for creators.

LinkedIn’s fight against AI slop is becoming a test of whether a professional network can protect useful human knowledge while allowing people to use AI as a legitimate writing aid. Success depends on the platform judging contributions fairly, even when AI helped shape the words.

Member reports will help LinkedIn refine the classifiers that decide which suggested and out-of-network posts get recommended. That gives readers’ impressions a role in improving the system even though research suggests people often confuse writing that feels inauthentic with writing produced by AI.

Members who rely on LinkedIn for professional knowledge should care because mass-produced posts and automated comments can crowd out the perspectives, ideas, and expertise they expect to find, making the feed less useful and harder to trust. Creators, especially freelancers and small teams, also need confidence that legitimate AI assistance will not cause their expertise to be treated as inauthentic.

An inaccurate judgment could affect whether a creator reaches people beyond an existing network, although LinkedIn has not explained when or how reports will change distribution. Smaller creators may be more vulnerable to authenticity warnings and any resulting loss of reach because they may rely on AI for support that larger organizations can obtain from employees or agencies.

LinkedIn now needs to decide how reports will be reviewed, how much influence they will have, and how creators will learn when a warning or classification affects their work. Clear standards, transparency about distribution, protection against malicious reporting, and a way to challenge mistakes will determine whether the system treats creators fairly.

In short: LinkedIn can reduce low-value automation without treating all AI-assisted work as suspicious. It needs a system that distinguishes genuine professional contributions from slop and corrects mistakes when that judgment fails.

A reporting system can only be as fair as the judgments feeding it, so members should ask whether a post is genuinely low-value or simply does not sound authentic to them.

TechCrunch: LinkedIn adds a button to report AI-generated ‘slop’
https://techcrunch.com/2026/07/30/linkedin-adds-a-button-to-report-ai-generated-slop/

LinkedIn: AI slop is a top priority for all of us
https://www.linkedin.com/posts/hsrinivasan1_ai-slop-is-a-top-priority-for-all-of-us-activity-7488612009660346374-xZiY/

arXiv: “That’s AI Slop, You Bot!” Studying Accusations, Evidence, and Credibility in Online Discourse Towards LLM-Generated Comments
https://arxiv.org/abs/2606.12073

The Substack Post: Against Claudefishing
https://post.substack.com/p/against-claudefishing

Substack Help Center: How can I detect AI on Substack?
https://support.substack.com/hc/en-us/articles/50891130623508-How-can-I-detect-AI-on-Substack

LinkedIn Help: Verifications on your LinkedIn profile
https://www.linkedin.com/help/linkedin/answer/a1359065

LinkedIn Help: Request verification for your LinkedIn Page
https://www.linkedin.com/help/linkedin/answer/a7179108

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing support, AEO/GEO/SEO optimization, image concept development, and editorial structuring support from ChatGPT, an AI assistant. All final editorial decisions, perspectives, and publishing choices were made by Alicia Shapiro.