LinkedIn is testing a reporting option that lets members flag posts as seeming like AI slop. When used, the report privately signals through the author’s analytics dashboard that others found the content inauthentic or overly reliant on artificial generation. Flagged posts then receive reduced distribution in feeds, treated much like material users have marked as not interesting.
The change sits alongside a quieter retreat from LinkedIn’s own generative tools. The prominent rewrite-with-AI button that once appeared on posts and messages is being removed. In its place comes a narrower proofreading feature that corrects grammar and spelling without rewriting the author’s voice. The company acknowledges that its earlier assistance features helped feed the very problem it now seeks to contain.
These steps follow an earlier effort to limit the reach of AI-generated material and further refinements to detection systems. The goal is fewer synthetic posts surfacing in recommendations. Yet the scale of the issue remains substantial. A recent analysis by the detection firm Pangram identified LinkedIn as the most AI-saturated major platform, with more than 40 percent of long-form posts classified as fully machine-generated. Beyond individual users, automated accounts have poured content onto the site. LinkedIn reports catching hundreds of thousands of bot-driven comments and blocking billions of attempted automated posts in recent months alone.
Professional networks have always balanced self-presentation with credibility. The arrival of capable language models accelerated a particular kind of low-effort signaling: polished but hollow updates, generic thought leadership, and comment threads that read as if assembled by template. LinkedIn’s earlier decision to embed rewriting tools made participation easier and, in the process, normalized the output. Rolling those tools back while adding a user-facing flag is an admission that the platform helped create the conditions it now wants to moderate.
Whether the combination of reporting, reduced reach, and lighter assistance will shift behavior is uncertain. Platforms have long struggled to contain spam, bot networks, and low-quality content once volume reaches a certain threshold. Detection improves, users adapt, and new generation methods appear. The private nature of the flag may encourage more reporting than public shaming would, yet it also limits social pressure on serial offenders. Reduced algorithmic visibility can slow the spread of flagged material without removing it, leaving the underlying posts still visible to those who seek them out.
LinkedIn’s latest adjustments therefore represent incremental containment rather than a decisive reset. They address both the supply of AI-assisted writing the company once encouraged and the downstream effects that followed. The professional feed remains a place where authenticity is claimed more often than demonstrated, and the tools now being tested will determine how much of that claim survives contact with the next wave of generated text.
