How Linkedin Is Fighting Back Against AI Slop and Comments

Have you seen a lot of AI slop on LinkedIn lately?

LinkedIn sees it too. Now the company is taking steps to solve the problem, unleashing new technology to eliminate low-quality AI content from users’ feeds.

“Content creation on the platform is up 14% year over year,” says Laura Lorenzetti, VP and Executive Editor of LinkedIn. “That makes sense, right? AI can really help people unlock content creation. But it also means that a lot of people can produce a lot of very low-quality content.”

That contrast creates a challenge for LinkedIn: Many professionals use AI in their daily workflows, including helping to turn their ideas into LinkedIn posts. So LinkedIn needed a way to differentiate the two — squashing AI-generated content that lacks original thought, while allowing higher value content to thrive even if its creator used AI.

To do this, LinkedIn is targeting three core areas:

1. Generic AI-written posts and comments
2. Automation tools used to create AI content
3. Attention-bait videos

Its first efforts are starting to roll out now. Here’s how LinkedIn is attacking all three AI problems.

Stopping the AI Slop

You’ve surely seen posts like this: They feature a graphic with a generic quote, accompanied by anodyne advice.

LinkedIn is frustrated by these posts, because they run against what the platform most wants — which is for people to post nuanced insights from their areas of expertise. Content creators also hate these posts, because they make it harder for legitimately valuable content to find its audience.

“The feed is more competitive,” Lorenzetti confirms. “Content creation is up. Those are both true, and I think the timing of that is very clearly at the moment that there was a rise in AI.”

So how can LinkedIn identify AI slop at scale? To do this, LinkedIn developed what Lorenzetti calls an “AI solving AI” approach.

The company built technology systems that it says distinguish between original thinking and posts that lack uniqueness or substance. The tech, built in partnership with LinkedIn’s editorial teams, can learn over time by identifying patterns in how members engage, as well as what language adds perspective, context, or expertise instead of just repeats existing ideas without contributing anything new.

Much of this starts by having human editors and content managers annotate thousands of posts, labeling them as either generic or original based on detailed definitions of what constitutes low- or high-quality content. Typically, multiple people review each post to ensure consistency.

These human-labeled examples then train machine learning models that can identify patterns in content at scale.

Stopping the AI Bots

If you post on LinkedIn, you’ve surely seen this problem too: Users comment on your posts, but their comments are suspiciously formal, repetitively formatted, or simply summarize and repeat what your post was about.

These are clear signs of AI comments, facilitated by a host of new AI products that automate commenting.

Why would someone want an AI tool to comment on their behalf? There are many reasons. Comments can often gain great visibility on the platform. And if a post gets a lot of comments, the LinkedIn algorithm is more likely to believe that the post is popular and worth sharing with others.

LinkedIn had already technically banned these AI commenting tools; they are against the platform’s terms of service. But users have been using them anyway.

To solve this problem, LinkedIn is doing something similar to what it does with generic posts. The company is building classifiers to identify low-quality AI comments — looking at the actual language in the comments, and also the patterns and volume in which comments are posted. (If someone’s using an AI tool, for example, they might comment much faster and more often than a regular user.)

Stopping the Attention-Bait Videos

LinkedIn is also targeting what it calls “attention-bait videos” — content designed purely to keep people watching without adding real value.

For example, you might see a lengthy video of construction accidents paired with generic workplace safety advice. Or someone might post extended footage of a manufacturing processes accompanied by vague business platitudes.

“It’s doing what AI slop is doing but in a much more visual way,” Lorenzetti says. These videos were often proven on other platforms like Instagram, where they performed well. Now users are posting them on LinkedIn just to grab attention.

When You’ll See Changes

LinkedIn won’t delete posts it believes are AI slop. Instead, its algorithm will depress that content’s reach — often not allowing it to be seen beyond a poster’s first-degree contacts.

The company is taking a gradual approach to implementation. Lorenzetti says it may take several months to fully impact user experience, as the company continues to refine its detection systems and monitor for new forms of problematic content.

And LinkedIn has no illusions: Even as it fights this battle, other problems will emerge — as people find new and creative ways to game its algorithm.

“AI slop is just the latest problem,” Lorenzetti says. “We’ll keep paying attention.”

Have you seen a lot of AI slop on LinkedIn lately?

LinkedIn sees it too. Now the company is taking steps to solve the problem, unleashing new technology to eliminate low-quality AI content from users’ feeds.

“Content creation on the platform is up 14% year over year,” says Laura Lorenzetti, VP and Executive Editor of LinkedIn. “That makes sense, right? AI can really help people unlock content creation. But it also means that a lot of people can produce a lot of very low-quality content.”