LinkedIn is implementing a user-driven flagging system to identify artificial intelligence-generated "slop" on its professional network. The update, announced by Chief Product Officer Hari Srinivasan,allows members to notify authors when their content feels inauthentic.
The 62% AI-content share that alarmed Pangram
The urgency behind LinkedIn's new tool stems from startling data provided by the AI detection startup Pangram. According to the report, LinkedIn accounted for approximately 62% of all AI-generated content flagged by Pangram, a disproportionate figure considering the platform represented only one-third of the total items scanned. This suggests a systemic infestation of synthetic text within the professional sphere.
The scale of the issue extends beyond short updates. pangram's data suggests that as much as 40% of long-form text on LinkedIn is AI-generated, the highest percentage among popular social platforms. This surge in "slop"—a term used to describe low-quality, generative AI content—threatens to erode the trust and authenticity that underpin professional networking.
How Hari Srinivasan's 'Seems like AI slop' button works
To combat this trend, LinkedIn has introduced a specific button labeled "Seems like AI slop," located within the three-dot menu of individual posts. When a user activates this flag, the author of the post receives a private notification via their analytics dashboard. chief Product Officer Hari Srinivasan stated that this human-centric feedback loop helps LinkedIn tune its internal models and improve the overall quality of the user feed .
As reported in the source, LinkedIn has already blocked billions of AI-generated comments and posts in recent months . however, Srinivasan believes that human intuition is superior to automated detection,arguing that members should receive feedback from real people on what sounds authentic rather than relying on AI detectors that may produce incorrect results.
Chris Best and the Substack push against LinkedIn-style content
The struggle against generative noise is becoming a defining battle for social platforms. Substack CEO Chris Best recently highlighted this tension, announcing that Substack would integrate Pangram's detection tools specifically to avoid the pitfalls seen on other networks . Best explicitly stated that he did not want Substack to "turn into LinkedIn," referencing the perceived saturation of AI-generated filler on the professional site.
This industry-wide shift reflects a growing backlash against the "efficiency" of generative AI. While AI tools were initially marketed as productivity boosters, the resulting flood of homogenized content has created a "race to the bottom" in terms of quality, forcing platforms like LinkedIn and Substack to pivot toward authenticity-first policies.
The risk of flagging em dashes and proper grammar as AI
Despite the rollout of the flagging button, significant questions remain regarding the accuracy of human judgment. One primary concern is that users may mistake high-quality writing for AI generation; common markers of professional prose, such as the use of colons, em dashes, and strict adherence to grammar, are often the same signals that trigger AI detectors.
Furthermore, the report notes that many AI detection tools are themselves powered by AI, leaving their reasoning opaque and prone to false positives. It remains unclear how LinkedIn will handle disputes where a human author is flagged as a bot,or whether a high volume of "slop" flags will lead to algorithmic shadow-banning of legitimate professional voices.
Replacing 'enhance your post' with human-centric proofreading
In a notable pivot, LinkedIn is retiring its "enhance your post" feature, which used AI to rewrite user content. The company is replacing this with a proofreading tool designed to preserve the original voice of the user. This move suggests that LinkedIn recognizes its own internal AI tools may have contributed to the very "slop" it is now trying to eliminate.
By shifting from generative enhancement to simple correction, LinkedIn aims to restore a sense of human presence to the platform . The goal, as promised by Hari Srinivasan, is a visible reduction in synthetic noise and a return to genuine professional interaction.
Comments 0