For the past few years LinkedIn offered users a generative tool for rewriting posts. Now it is removing it. On 30 July 2026 it launched a “Seems like AI slop” button, with which members can report a post that seems machine-made, and the platform then reduces its reach.
The button under the three dots
The new option is hidden under the three-dot menu in the top right corner of a post. When a member flags a post, LinkedIn hides it from that reader, and the signal is used to train classifiers meant to catch slop before it reaches recommendations beyond the author's own followers. The author learns about the flag because LinkedIn notifies them privately in the analytics overview when members perceive their content as inauthentic or heavily marked by AI.
Hari Srinivasan, LinkedIn's Chief Product Officer, introduced the launch with the words “AI slop is a top priority for all of us. We really care about this. People come to LinkedIn to connect with real people and share their real perspectives, ideas, and expertise.”
The end of enhance your post
The announcement also included a turnaround on a tool that LinkedIn itself offered. It is removing the “enhance your post” feature, which could rewrite text with generative AI, and replacing it with a tool that “proofreads your words, but does not change your voice”. Proofreading stays, generating new content ends.
Srinivasan explicitly separates AI assistance from slop. He wrote that “AI and slop are not the same thing”. The target is low-quality text without its own perspective, which floods the feed and dilutes the reason people come to a professional network.
Pangram data behind the launch
The figures behind this step come from Pangram Labs, which specialises in detecting AI text. In a report from 9 July 2026, it scanned 1,002,627 posts across five platforms, namely LinkedIn, Medium, Substack, X and Reddit, excluding posts shorter than 50 words. For long-form posts above 250 words, more than 40% on LinkedIn came out as fully AI-generated.
Even more telling is the share of the whole. LinkedIn accounted for roughly a third of the scanned content, but it accounted for 62% of all flagged AI text, which, according to Pangram, makes it the platform with the highest share of AI content in the group. Across all platforms, one of the four longest posts was fully generated.
Automated comments and bots
Reporting is only part of the measures. According to Srinivasan, LinkedIn catches “hundreds of thousands of automated comment attempts” every day and in recent months has “blocked billions of other automation attempts”. It is expanding verification of profiles and pages and adding the option to block comments from specific company pages.
The pressure comes from the wider web. Substack deployed AI detection through a partnership with Pangram, and Cloudflare reports that automated traffic on the internet has overtaken human traffic. Platforms whose value rests on there being a human behind the content are starting to treat slop as an existential problem.
The risk of misusing the button
The new mechanism has a weakness. Tech ghostwriter Colin Steele warned that the button is “ripe for abuse”, meaning that companies and their employees can report competitors even without good reason. AI detection remains imprecise, and LinkedIn is building it on reports from people precisely because it admits the limits of an automatic detector.
From August the AI Act only requires labelling
LinkedIn is acting in the same week that EU regulation on the same topic comes into force. From 2 August 2026, Article 50 of the EU AI Act applies, which requires AI-generated content to be labelled in a machine-readable way. The regulation asks for transparency about origin. LinkedIn goes further and penalises low-quality AI text with reduced reach, whether or not it is labelled.
For European creators and brands building authority on LinkedIn, the calculation changes. The concrete impact of the regulation on an individual creator will depend on interpretation and enforcement, which are still settling. The platform's position is clear already, however. On a network where much long-form content is written by a machine, LinkedIn protects content with human input, because that is what sets it apart.
Comments