LinkedIn has launched a crowdsourced tool against the flood of machine-written posts filling its feed, adding a user-facing reporting button labeled “Seems like AI slop” and replacing the platform’s own AI post-writing tool with a proofreading feature. The moves were announced July 30, 2026. They mark one of the most direct acknowledgments by a major professional network that generative AI has created a content-quality problem.
What the new button does
The professional networking platform added a “Seems like AI slop” button, which appears when a user clicks on the options button on a post, and helps users flag posts they believe are low-quality or overreliant on AI. A LinkedIn spokesperson confirmed the button is being added to the ellipsis menu on LinkedIn posts and will be available on all posts and comments.
In a post on LinkedIn, the company’s chief product officer, Hari Srinivasan, acknowledged the Microsoft-owned social network is facing the problem. “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 said LinkedIn is “ramping up a series of new and improved classifiers that identify if a post is AI-slop or generally low-quality content,” with the company saying it will use that feedback to determine how much reach a post will get outside of the poster’s network.
Reports feed detection and moderation: they act as a signal that lets LinkedIn tune its AI models to better identify slop, with every flag acting as a labeled training example.
Scale of the problem
A study by AI text detection firm Pangram found 41% of long-form LinkedIn posts are “flagged as fully AI-generated,” as is 30% of short-form content on LinkedIn.
LinkedIn is introducing new classifiers to identify if a post is AI slop or other low-quality content, which would reduce the amount of slop appearing in suggested content recommendations from outside a user’s network. The crackdown extends to comments too. LinkedIn is suppressing generic posts from recommendations rather than removing them outright.
Srinivasan said the company detects and blocks hundreds of thousands of automated slop comment attempts every day, and has prevented “billions of other automation attempts (posting at scale, slop) in the last couple months alone.” The platform is also going after automation tools that generate AI content at scale.
LinkedIn replaces its own AI writing tool
The company is also removing its “enhance your post” tool, replacing it with a feature that proofreads posts without changing a user’s voice, while reinforcing profile and page verifications.
LinkedIn plans to test a feature within user analytics dashboards that flags when their own content may be coming across as inauthentic or heavily AI-written. Srinivasan said “AI and slop are not the same thing,” noting that “many people refine thoughts with AI” and that the platform wants users “to get feedback from real humans on what sounds authentic.”
LinkedIn has a problem it partially created for itself. After years of integrating AI writing tools directly into its product, the platform is now working to contain the flood of low-quality AI-generated content those tools helped produce.
The tension is a recurring theme across the industry: platforms that promoted generative AI features are now building filters to limit their most visible downsides. That dynamic is broader than any one network, and Microsoft Copilot also added new guardrails around its agentic AI tier in its own recent platform overhaul.
Broader platform movement
The move reflects a broader shift across online publishing platforms to cut back on AI content, as people have grown frustrated with the computer-written, inauthentic posts filling the web. Last week, newsletter platform Substack added a tool to help users identify when content they are reading was written by AI, through a partnership with Pangram.
LinkedIn and other platforms are stepping up efforts to rein in low-quality AI content without banning the technology entirely. The platform has not said when the full rollout will be complete. What is clear is that the professional network, which built its content ecosystem on the premise of authentic expertise, is now betting that user-generated signals can train classifiers faster than bad actors can game them, a wager that other major social platforms are watching closely.
