News

LinkedIn Wants You to Use AI Without Sounding Like AI

• By John K. Waters
• 08/28/2026

LinkedIn is trying to solve a problem that increasingly confronts every platform built around human conversation: What happens when artificial intelligence becomes good enough to help everyone write, but so ubiquitous that it makes everyone sound the same?

The Microsoft-owned professional network is not banning AI-generated writing. In fact, LinkedIn explicitly says using AI to help write a post is acceptable. Instead, it is trying to identify and reduce the reach of AI slop: low-effort content that may be polished but lacks an identifiable perspective or substantive value.

That distinction is becoming increasingly important as generative AI moves from novelty to routine writing tool.

LinkedIn introduced a "Seems like AI slop" feedback option on July 30, allowing members to flag posts and comments they believe fit that description. More than 1 million members used the option in its first few weeks, LinkedIn Chief Product Officer Hari Srinivasan wrote in a post on the platform.

At the same time, LinkedIn has been deploying classifiers designed to identify AI slop and other low-quality material. Srinivasan said members are now seeing about 40% fewer views of content LinkedIn classifies as slop compared with a few weeks earlier. LinkedIn has clarified that the reduction is due to its classifiers, rather than simply to users pressing the new feedback button.

The result is something more nuanced than a crackdown on AI writing. LinkedIn is effectively trying to distinguish between AI that helps people communicate and AI that replaces the person doing the communicating.

AI Is Not the Problem
LinkedIn has been unusually direct on that point.

"It's ok to use AI to help you write," the company said in a June explanation of its approach, adding that posts and comments still need to represent the member's own voice and perspective.

LinkedIn's existing guidance similarly describes AI as a tool that can augment a person's expression. It recommends that members review and edit AI-assisted material before publishing it and disclose substantial AI involvement when it would not otherwise be apparent.

Srinivasan made the distinction even more explicitly when announcing the latest changes: "AI and slop are not the same thing." He said many people use AI to refine their thoughts, while LinkedIn wants to help users recognize when the result comes across as inauthentic.

That leaves LinkedIn with a difficult technical and editorial problem.

The platform cannot simply identify AI-generated text and suppress it because LinkedIn allows people to use AI to help them write. Instead, it must determine whether a post contributes perspective, context, or expertise, or merely reproduces the smooth, generic language that generative AI systems can produce at enormous scale.

LinkedIn said an early version of its system correctly identified generic content 94% of the time in testing. When content appears AI-generated and lacks a clear perspective, the company says it is less likely to be distributed widely beyond the author's immediate network.

The Scale of the Problem
There is evidence explaining LinkedIn's urgency.

AI detection company Pangram Labs analyzed browsing data collected via its Chrome extension and estimated that up to 41 percent of long-form LinkedIn content users encountered was likely fully AI-generated. The finding, first reported by 404 Media, relies on automated AI detection and should therefore be treated as an estimate rather than a definitive measurement of LinkedIn's content.

LinkedIn is also fighting automation beyond the posts themselves. Srinivasan said the company is catching hundreds of thousands of automated comment attempts each day and has blocked billions of other automated attempts over several months.

The company had already announced changes in March aimed at automated comments, engagement pods, recycled posts, and engagement bait. LinkedIn said it was upgrading its feed with generative recommenders and large language models while simultaneously trying to reduce content and interactions that appear automated or inauthentic.

There is some irony in that approach. LinkedIn is using increasingly sophisticated AI to decide which uses of AI make LinkedIn less useful.

When AI Makes You Sound Like AI
Perhaps the clearest indication of where LinkedIn is drawing the line is a change to its own generative AI tools. The company is removing its "enhance your post" feature, which could rewrite a user's post with AI. LinkedIn plans to replace it with a proofreading feature designed to correct users' wording without changing their voice.

That's a small product change with a larger implication.

LinkedIn appears to be concluding that the most useful role for generative AI in professional communication may not be to speak for people, but to help people speak more clearly for themselves.

The company is also testing private feedback for authors whose posts are perceived by other members as inauthentic or overly reliant on AI. Srinivasan said LinkedIn wants that judgment to incorporate feedback from real people rather than depending entirely on an AI detector. No single piece of member feedback determines how a post is distributed, according to the company.

That distinction matters because detecting AI-generated writing remains an imperfect science, particularly as human and machine writing become increasingly intertwined. The problem is no longer simply determining whether AI touched a piece of text. A person might write a post, ask an AI assistant to tighten it, rewrite a paragraph, restore some original wording, and publish the result. LinkedIn itself endorses versions of that workflow.

The harder question is whether anything recognizably human survived the process. For LinkedIn, the answer increasingly appears to determine how far a post travels. The company's emerging policy is therefore less a prohibition on artificial intelligence than an attempt to preserve scarcity in something AI cannot mass-produce nearly as easily: individual experience and perspective.

AI can help you write the post. LinkedIn just wants the post to still sound like you.