LinkedIn suppresses AI-generated 'slop' from recommendations in feed overhaul

LinkedIn has introduced new algorithmic measures to reduce the visibility of posts identified as AI-generated "slop," targeting content that lacks authenticity, originality, or substantive value. The changes, detailed in a blog post by VP of Product Laura Lorenzetti, aim to suppress engagement bait, recycled "thought leadership," and generic material from user recommendations. While these posts remain accessible to the author’s direct connections and followers, they will no longer appear in the broader feed of other users.

The crackdown specifically targets posts exhibiting obvious signs of artificial construction, such as the repetitive "it's not X, it's Y" phrasing structure. To identify such content, LinkedIn engineers collaborated with its in-house editorial team to analyse engagement patterns. This partnership sought to distinguish between content that adds perspective, context, or expertise, and material that merely repeats existing ideas without contributing new insights. The platform has not disclosed the specific technical methods used to define or detect this "slop," nor how it will handle edge cases where human writing mimics AI patterns.

This policy shift follows a recent period of significant user backlash regarding "em dash discourse," a trend where members debated the use of em dashes as a marker of AI-generated text. The debate highlighted concerns over stylistic markers in large-language models, which are trained on human-written content that frequently employs such punctuation. The influx of AI-generated material has exacerbated LinkedIn’s historical struggles with shameless self-promotion and borderline spam, problems that were already prevalent on the professional network prior to the rise of generative AI.

Despite the crackdown, LinkedIn maintains a distinction between AI-generated slop and "AI-assisted" content, which remains permitted provided it encourages meaningful conversation or offers original ideas. The company continues to promote its own generative AI tools, including a "rewrite with AI" button within the post composer, creating a tension between suppressing AI output and encouraging its use. This approach underscores the platform's effort to balance the utility of AI tools with the need for authentic professional discourse.

Initial results from these efforts have been described as "encouraging," with LinkedIn expecting further declines in AI slop in the coming weeks and months. The move positions LinkedIn to address the quality of its feed more aggressively than many other digital platforms, which are also grappling with the influx of automated content. As the professional network refines its curation algorithms, the long-term impact on user engagement and content quality remains to be seen.