LinkedIn will suppress AI-generated posts without deleting them.

The LinkedIn feed has become unreadable, and if you spend not even five minutes a day on it, you don't need someone to draw it for you. Interchangeable 'leadership' carousels, hooks that all start with 'It's not X, it's Y', and comments that rephrase the post by changing an adjective. The platform has decided to tackle the problem.

How LinkedIn plans to sort

The system detects generic posts based on stylistic and behavioral signals: uniformly smooth vocabulary, absence of personal viewpoint, repetitive structures typical of language models, artificial engagement. The platform claims 94% accuracy in its initial tests. A good number, except that no one knows the false positive rate. How many legitimate posts will be mistakenly buried remains a mystery that LinkedIn hasn't deemed useful to clarify.

Detected posts will not disappear; they will remain visible to the author's direct contacts. However, they will no longer appear in others' feeds. In other words, your copied post from ChatGPT will be read by your colleagues and former classmates, no one else. Full deployment will take "several months," giving those used to copying a little time to learn how to rephrase.

Laura Lorenzetti, Vice President of Product, clarifies that using AI as a writing aid remains acceptable. It is the content that contains "neither viewpoint, nor expertise, nor experience" that will be targeted. The boundary between the two is clearly evident for an algorithm.

The paradox that LinkedIn prefers not to comment on

In November 2025, LinkedIn activated the training of its AI models on the data of its 34 million French members without prior consent. The "Rewrite with AI" button is always offered to all Premium subscribers. And its own CEO admitted in mid-2025 that the AI tools on the platform were "not very popular." A few months later, the result is here: more than half of long English posts would be generated by artificial intelligence, and these posts average 45% less engagement than human content.

LinkedIn is fighting a problem it partly created, with a detector whose limitations it does not share, on a timeline it does not specify. All on a platform with 17.4 million monthly visitors in France (fifth-largest global market), where every drop in reach translates into missed professional opportunities.

The good news is that your feed should eventually improve. The downside is that if your own posts are affected, you'll be the last to know.

👉🏻 Follow the latest tech news in real time: add 01net to your Google sources, and subscribe to our WhatsApp channel.

Opera One - Web Browser Boosted by AI

By: Opera

Download for free

Source : Laura Lorenzetti/LinkedIn