LinkedIn's 'Seems Like AI Slop' Reporting Button

The mechanic itself is simple: a 'Seems like AI slop' option now sits in the three-dot menu on every post, and clicking it hides that post from the reporter's feed while the report is fed back into LinkedIn's AI-detection classifiers ^[1][2]. It is not a moderation action against the poster, more like a supercharged 'not interested' - flagged posts get reduced algorithmic distribution, and the original poster gets a private nudge in their analytics dashboard suggesting their writing may have read as inauthentic or AI-heavy ^[3][4]. LinkedIn paired the launch with a second, quieter change: retiring its 'enhance your post' AI rewrite tool and replacing it with a proofreading feature that fixes grammar without rewriting a user's voice ^[3].

The timing is not a coincidence. Independent detection vendor Pangram scanned over a million posts across LinkedIn, X, Reddit, Substack and Medium and found LinkedIn accounted for roughly one-third of all posts scanned but 62% of all AI content flagged, with over 40% of its long-form posts fully AI-generated ^[5][6]. Originality.ai's tracking tells an even starker story: from a 2024 baseline of 54% likely-AI long-form posts, its most recent sample of 5,000 July 2026 posts found 81.2% likely-AI ^[7]. Trade press had already flagged LinkedIn and X as the worst offenders for AI slop weeks before the button shipped ^[8], which suggests LinkedIn built this feature reactively, under public pressure from detection-company research rather than as a proactive quality initiative.

The most cutting read on this launch, echoed widely in early user reaction, is that LinkedIn has little structural incentive to actually eliminate AI slop. Every flag is a free, human-labeled training example for the classifier that will eventually decide what counts as 'AI' on the platform - a crowdsourced dataset LinkedIn would otherwise have to pay to build. It is also worth noting that LinkedIn is the party that helped create this problem in the first place: its own 'enhance your post' rewrite tool actively encouraged AI-assisted posting before being quietly retired in favor of a proofreader that promises not to alter a user's voice ^[3][9]. That reversal amounts to LinkedIn undoing a feature it once promoted, now framed as a correction rather than an admission.

Coverage of the launch leaned into this irony directly, with one outlet describing LinkedIn's users as having been 'bathing in AI slop' for years before the company 'offers a shower' ^[10]. The skepticism isn't just about motive - it's about the business model. LinkedIn's engagement metrics and its owner Microsoft's broader AI ambitions both benefit from a platform full of AI-adjacent content and behavioral data, so a lightweight, crowdsourced, non-punitive reporting button is a low-cost way to look responsive without meaningfully changing what gets amplified.

LinkedIn's own framing acknowledges the core problem: 'slop' is subjective. Srinivasan has said the definition 'is hard to define and the definition changes,' which is precisely why the system leans on user reports rather than a fixed rule ^[1]. That fuzziness is also where the backlash concentrates. Ghostwriter Colin Steele called the button 'ripe for abuse,' warning that a crowdsourced, unverified reporting tool can be weaponized by competitors to suppress legitimate content rather than genuine slop ^[4]. The risk compounds because 'AI-assisted' and 'AI-generated garbage' are not the same category, yet the flag button collapses both into one gesture - a polished, professional post refined with AI could get the same treatment as a templated engagement-bait post with no original thought behind it.

LinkedIn's counter is explicit: 'AI and slop are not the same thing. Many people refine thoughts with AI, and we believe they want to know when they sound inauthentic,' Srinivasan said, trying to draw a line between assistance and inauthenticity ^[4]. Substack's Chris Best offered an outside-in verdict on where that line matters most, saying his platform doesn't 'want to wait until your Substack app turns into LinkedIn' ^[4]- treating LinkedIn's current state as a cautionary tale other platforms are actively designing around, rather than a problem LinkedIn has already solved.

A flag button addresses demand-side friction - annoyed readers - but does little about the supply side that produces the slop in the first place. Separate engagement data suggests the market was already partially self-correcting: AI-generated LinkedIn posts get roughly 45% less engagement on average than human-authored ones ^[11], meaning much of the audience already discounts obviously templated content without needing a reporting tool. That raises the real question the button doesn't answer: who is still posting slop despite the engagement penalty, and why?

The answer traces to a templated content economy built specifically to game LinkedIn's algorithm - generic 'thought leadership' formats, manufactured personal-narrative hooks, and recycled engagement-bait structures produced at scale by small AI content operations rather than individual users experimenting with a writing assistant. A flag button that hides one post at a time, trains a classifier slowly, and relies on users noticing and caring enough to click, is a mismatch in speed and scale against a supply chain built to produce that content continuously. Until LinkedIn addresses the production side rather than just the consumption side, the reporting button functions more as a sentiment valve for frustrated users than a structural fix.