LinkedIn is moving to combat the flood of low-quality, AI-generated content—often called "AI slop"—that has been cluttering its feed. On Thursday, the company introduced a new feature allowing users to click a "seems like AI slop" button on posts they suspect were written by artificial intelligence. This initiative is part of a wider trend across online platforms to reduce AI-generated content, as users grow increasingly frustrated with inauthentic, computer-written posts dominating the web. For example, last week, the newsletter platform Substack partnered with Pangram to add a tool that helps users identify AI-written content. Meanwhile, Pangram announced $9 million in new funding this week to address the issue of AI content flooding the internet. The problem also affects emerging startups; Digg shut down its Reddit competitor in March, citing an inability to control the number of bots on its site. Internet infrastructure firm Cloudflare reports that the situation is worsening, with bot traffic now surpassing human-generated requests online—a milestone reached faster than previously anticipated.

In a LinkedIn post, Chief Product Officer Hari Srinivasan acknowledged that the Microsoft-owned social network is facing similar challenges. "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," he said. Srinivasan explained that the new "Seems like AI slop" button is one of several measures LinkedIn is implementing to curb low-quality, AI-generated content. The company is also strengthening its automation defenses, now blocking "hundreds of thousands" of automated comment attempts daily, and millions more over the past couple of months. Additionally, Srinivasan noted that LinkedIn is introducing new classifiers to identify AI slop or other low-quality content, which will reduce the amount of such content shown in suggested recommendations from outside users' networks. The button ties into this effort by providing signal data that helps LinkedIn fine-tune its AI models to better detect slop.

Furthermore, LinkedIn will begin privately notifying users in their dashboards when others perceive their content as inauthentic due to heavy AI use. The company hopes this will encourage posters to improve their writing, especially if they are using AI to refine their own work rather than creating full-on "slop" content. Notably, LinkedIn is removing its own "enhance your post" feature, which used AI to help users write, and replacing it with a proofreading tool that preserves the user's voice instead of altering it. Other updates include expanding access to profile and page verification tools, as well as adding an option to block comments from company pages users no longer wish to see, Srinivasan said.