LinkedIn announced a suite of anti-AI-slop measures on July 30, 2026, headlined by a new user-facing report button that lets members flag posts they suspect were generated by AI. The button, labeled "seems like AI slop," appears in the three-dots menu on individual posts. When clicked, LinkedIn hides the post from the reporting user and displays a thank-you message for the feedback (The Verge).

The announcement came from LinkedIn Chief Product Officer Hari Srinivasan, who posted the details on LinkedIn's own platform. "AI slop is a top priority for all of us. We really care about this," Srinivasan wrote. "People come to LinkedIn to connect with real people and share their real perspectives, ideas, and expertise" (TechCrunch).

The scale of the problem is measurable. According to AI detector Pangram, 41 percent of longform LinkedIn posts were flagged as completely AI-generated (The Verge). LinkedIn's own enforcement data backs up the concern: the platform blocks hundreds of thousands of automated comment attempts daily as of July 2026, and blocked millions of automation attempts in the months preceding the announcement (TechCrunch).

The report button serves a dual purpose. Beyond removing the post from the individual user's feed, each report feeds a signal back into LinkedIn's content classification models, helping the platform tune its automated detection of low-quality or AI-generated content (The Verge). LinkedIn is also introducing new classifiers designed to identify AI slop and reduce its appearance in suggested content surfaced from outside a user's network (TechCrunch).

The company is pairing automated detection with a nudge toward creators themselves. LinkedIn will begin privately flagging in users' dashboards when their own content is perceived as inauthentic due to heavy AI use (TechCrunch). The mechanism is reminiscent of the engagement-quality signals platforms have long given advertisers, except here the feedback targets the content creator's own output.

LinkedIn is also retiring its existing "enhance your post" AI-writing feature, which presumably contributed to the slop problem from the supply side. In its place, the company is rolling out a tool that proofreads users' words without altering their voice (TechCrunch). The distinction matters: the old feature rewrote posts, while the replacement constrains itself to surface-level correction.

Additional measures include expanded access to profile and page verification tools, and a new option for users to block comments from company pages they no longer wish to engage with (TechCrunch).

Looking at the broader context, LinkedIn's move is the clearest acknowledgment yet from a major professional network that generative AI has shifted from novelty to nuisance on its platform. The 41 percent figure from Pangram is the kind of threshold that forces a platform's hand; when roughly two in five longform posts may be machine-generated, the core value proposition of a professional network built on authentic expertise is directly undercut.

The architecture LinkedIn is assembling is worth noting for those who build or moderate content platforms. Rather than relying on a single detection vector, the company is layering three approaches: user-generated reports as a labeled-signal source for model training, automated classifiers that act at the recommendation layer to suppress slop before it reaches users outside a creator's network, and creator-facing dashboard flags that introduce a feedback loop on the production side. This is a human-in-the-loop detection pipeline with explicit model-tuning intent baked into the UI.

The retirement of "enhance your post" is the quietest but arguably most consequential part of the announcement. Platforms that ship AI-writing tools and then build anti-AI-slop infrastructure around them are working at cross-purposes. Retiring the generative feature in favor of a proofreading-only replacement eliminates one internal source of the problem LinkedIn is now spending engineering effort to combat.

The dashboard flagging mechanism raises questions LinkedIn has not yet answered publicly. What threshold of AI-assisted language triggers the flag? Is the detection model the same one powering the recommendation-layer classifiers, or a separate system? And does the flag carry any downstream consequence for a post's reach beyond the nudge? These are the details that will determine whether the measure functions as gentle encouragement or as a soft penalty that shapes creator behavior through distribution effects.

For now, LinkedIn has placed the tools and signals in front of users. The effectiveness will depend on classifier accuracy, a problem the AI-detection field has not fully solved anywhere. But the combination of user reports as training signal, recommendation-layer suppression, and creator-facing feedback gives the platform multiple intervention points, which is more than most social networks have committed to.