After allowing, and even encouraging, the use of AI to create content on their platforms, the web giants are now backpedaling to reduce the flood of AI-generated posts. We take stock of this ironic situation.
A few weeks ago, LinkedIn launched a new feature to flag posts that appear to have been generated by AI. In the dropdown menu of a post, there is now an option reading “Looks like AI slop.” The American platform seemed to be responding to a recent Panagram study estimating that 41% of long posts and 30% of short posts on LinkedIn were generated with AI.
When unveiling the “anti-slop” button, Hari Srinivasan, VP of Product at LinkedIn Ecosystem, firmly positioned LinkedIn in the fight against bad AI-generated content.
“‘AI slop’ is a top priority for all of us,” he said in a LinkedIn post. “It’s a topic we genuinely care about. People come to LinkedIn to connect with real people and share their true perspectives, ideas and expertise.”
LinkedIn is currently developing tools to detect and demote content that appears inauthentic in the news feed. It has also pulled back an AI-assisted writing option.
“We asked ourselves why users were publishing messages using AI,” Hari Srinivasan wrote in the same post. “The answer is simple: LinkedIn isn’t a social network where you settle for a single word, and users feel more confident when they run their posts through AI. Building on that insight, we are removing the ‘Improve your post’ feature that appears when you write a post or message, and replacing it with a feature that proofreads your text without changing your style.”
Reddit’s positioning
LinkedIn is not alone in taking action. Meta has added automatic labels to alert users to posts that use image-generation technologies (C2PA, IPTC). Recently, TikTok added a slider that lets users choose the level of AI-generated content they see.
“The goal is to please users by offering them content that is more and more engaging and makes them spend more and more time on the platform,” explained Alexandre Gravel, cofounder of the agency Toast, in a recent post.
Reddit, for its part, has a different concern.
“Reddit’s pride, and its business model, depend on the humanity and authenticity you find on the platform,” he continues. “Why? Because Reddit sells its content to Google and OpenAI. That allows those engines [Google and OpenAI] to somewhat detect human intentions. If a flood [of AI-generated content] starts to settle on the platform, it causes big problems.”
As a result, Reddit has chosen to fight AI-generated content fiercely, banning it entirely from its platform.
The party isn’t over yet
Among the giants, the battle is therefore on. Each is going about it in its own way, according to its interests and its level of tolerance. As a content creator, should you start worrying now? Should you turn away from AI and go back to handcrafted content-writing techniques? Is it the end of the free ride for content creators? Not yet, says Francis Jette, consultant and trainer at Isarta.
“Platforms are starting to move, but I don’t think we can declare the party over. We are in a period of evolution and adaptation. The web giants are assessing how they will adapt to artificial intelligence content.”
He points out that, quite recently, AI-generated content went viral on Instagram while claiming to be “handmade.”
“To avoid being flagged, the goal shouldn’t be to reduce your use of AI, but rather to learn how to use artificial intelligence well, so that algorithms and human beings, ultimately, don’t perceive that the content was produced by AI,” he adds.
Rather than getting bogged down in a “pro” AI or “anti” AI stance, Francis Jette invites content creators to return to the fundamentals of brand content.
“The important question to ask is: how do we create content that is unique to our brand, that demonstrates our expertise, our field experience, our authority, while building trust? It’s by relying on our expertise that we manage to create content that is true and authentic.”
training.isarta.com 
