
LinkedIn Targets AI-Generated Comments to Protect Organic Reach
LinkedIn is extending its fight against AI slop to comments: generic replies lose visibility, and creators now receive notifications only for quality comments. The change affects anyone building organic reach on the platform.
Blog du Modérateur reports that LinkedIn is extending its fight against AI-generated content to the comment layer. According to the outlet, generic replies lose visibility, and creators no longer receive notifications for every comment, only for those the platform deems worthwhile. The report does not specify how quality is assessed, which signals carry weight, nor whether the change is already live for every account.
For teams that treat LinkedIn as an acquisition channel, this is not a cosmetic tweak. Comments have long served as a cheap way to borrow reach: a short reply under a popular post puts a name in front of an audience that did not follow the account. If low-effort replies are demoted, that arbitrage narrows.
What actually changes for reach
The announcement points to two levers. The first is visibility: a comment that reads as filler is less likely to be shown to others. The second is notification: creators are spared a stream of responses that add nothing. Together they suggest the platform is treating the comment thread as a surface worth curating, rather than a passive appendage to the post.
What the report leaves open matters. Whether demotion is automatic or review-based is unknown. Whether accounts that post genuine replies are rewarded with broader distribution is not stated. And whether the quality signal applies to the commenter's other activity on the platform is unspecified. Teams should avoid building a model on guesses and instead observe their own numbers.
The wider context is familiar. Since generative tools made text cheap, platforms have been steadily tightening the terms on which automated output earns distribution. Comments were one of the last places where volume still worked.
What a team can check this week
Start with what you already published. Review the comments your team left over the past month under third-party posts and ask a blunt question: would this have been worth writing if no one could see it? Replies that only restate the post, thank the author, or add a generic observation are the ones most exposed.
Then look at your own threads. If creators now receive fewer notifications, the comments that do surface matter more. A question that invites a specific answer from a specific person will outperform an open invitation to react. Practical steps:
- Audit recent third-party comments and cut those with no concrete detail, number or example.
- Rewrite your own call to action so it asks for a precise experience, not a general opinion.
- Track notification volume and comment impressions over the next few weeks, and treat early movement as directional rather than conclusive.
- Brief whoever manages the account that reply volume is no longer a metric worth reporting on its own.
The monetisation angle
For businesses whose funnel starts with organic conversation, the cost of a low-quality reply is now paid in reach rather than in reputation alone. That pushes teams toward fewer, denser contributions, and toward formats where substance is visible at a glance: a stated result, a counter-example, a named constraint.
It also raises a question the report does not answer: whether paid distribution is affected in any way. Nothing in the announcement suggests it, and it should not be assumed.
The direction is clear enough to act on. Comments are being read as content, and content is being judged on whether it says anything.
Also worth reading
-
Research-In-The-Loop: Keeping User Research Sync With AI-Speed Development
Nielsen Norman Group outlines a research-in-the-loop model, where research defines what each release must measure as AI accelerates development. The approach adapts user research to faster shipping cycles rather than slowing them down.
-
Instagram brings its Edits video app to Mac
Instagram has released a Mac version of Edits, its video editing app, aimed at longer projects and described as offering new exclusive features. The report gives no details on the feature list, the rollout schedule or pricing.
-
Giscard Punk, horreur euclidienne, techno-naturalisme : trois esthétiques web
L'ADN identifies three web aesthetics — Giscard Punk, Euclidean horror and techno-naturalism — that reflect how we relate to technology. For brands, the visual layer of a site is now part of how traffic is captured and kept.

