A study recently found that nearly half of all long-form posts on the platform are AI-generated.

LinkedIn has begun rolling out a new “Seems like AI slop” option that lets users flag posts they believe are low-quality, AI-generated content. The feature appears under the three-dot menu on posts and is part of the platform’s broader effort to reduce generic AI writing from users’ feeds. According to LinkedIn, flagged posts will be hidden from the reporting user, while the company uses the feedback to improve its systems for identifying inauthentic content.

The move comes as LinkedIn tries to combat a growing reputation for being flooded with repetitive AI-written “thought leadership” posts. The company says it’s expanding its detection systems to reduce the reach of generic AI-generated content and is replacing its previous AI writing assistant with a lighter proofreading tool designed to preserve a user’s own voice.

A recent report from AI detection startup Pangram Labs found LinkedIn was the most AI-saturated social platform it analyzed. The company estimated that 41% of long-form LinkedIn posts were entirely AI-generated, while the platform accounted for 62% of all AI-generated content detected across the dataset despite making up only about one-third of the posts analyzed. Pangram also found that nearly 30% of short LinkedIn posts were fully AI-generated. The company noted that LinkedIn’s built-in AI writing tools likely contributed to the trend, though it acknowledged AI detectors are not perfect and can produce false positives.

The problem extends beyond LinkedIn, but to a lesser extent. Pangram found that Twitter (X) had the highest overall share of AI involvement, with 23.9% of long-form posts classified as fully AI-generated and another 22.9% identified as AI-assisted. More than one in five Substack posts were also flagged as either fully AI-generated or AI-assisted.