LinkedIn fights 'AI slop' as study finds 40% of long posts are synthetic
The professional network is deploying new reporting tools and restricting AI writing features to combat a surge in low-quality generated content.
LinkedIn is intensifying its crackdown on generative AI "slop" after research revealed the platform has become a primary hub for synthetic long-form content. The move comes as the company seeks to protect the authenticity of professional networking against a tide of automated posts.
According to a study by AI detection startup Pangram, more than 40% of LinkedIn posts exceeding 250 words are fully AI-generated. Pangram's analysis, which spanned over one million posts across X, Reddit, Substack, and LinkedIn, identified the professional network as having the highest rate of AI saturation for long-form content among the platforms studied. In response, LinkedIn has introduced a "Seems like AI slop" reporting button within the three-dot menu of posts and is deploying improved classifiers to filter low-quality AI content from suggested feeds. The company also blocks hundreds of thousands of automated comment attempts every day.
The Shift Toward Human Voice
To further curb the proliferation of synthetic text, LinkedIn is retiring its "Enhance your post" AI writing tool. The company is replacing it with a more conservative proofreading feature specifically designed to preserve the user's original voice rather than rewriting content.
"AI slop is a top priority for all of us. We really care about this," said Hari Srinivasan, LinkedIn's Chief Product Officer, emphasizing the urgency of the cleanup.
Industry-Wide Pollution
This effort is part of a broader industry struggle against mass-produced, low-quality AI content. Other platforms are adopting similar transparency measures; for instance, Substack integrated Pangram's detection tools on July 21, allowing readers to view the human-versus-AI split of a piece of content. Substack CEO Chris Best noted that when content "made by no one" dominates human spaces, it "pollutes the commons" and obscures genuine human voices.
The Stakes for Professional Trust
For LinkedIn, the stakes are existential. The platform's core value depends on authentic human connection and professional trust. If the feed becomes dominated by artificial content, the utility of the site for networking and discovery erodes, risking a migration of users toward platforms that can more reliably verify human authorship.
What to Watch
As LinkedIn shifts from AI-assisted creation to AI-assisted proofreading, the industry will be watching to see if user-led reporting and stricter classifiers can actually move the needle on the 40% saturation rate. The effectiveness of these tools will determine whether professional social media can remain a space for human expertise or succumb to the noise of the generative era.