LinkedIn's Fight Against AI Slop

I think it's awesome that LinkedIn is taking such a public stance on AI slop.

Tim Jurka, LinkedIn's VP of Engineering, published the technology behind their fight against AI slop recently. LinkedIn is planting the "human trust" flag firmly in the ground, and as a platform that prides itself on creating economic opportunity for every professional on earth, they have to take this seriously.

So, LinkedIn's classifiers now screen every post distributed beyond your immediate network at 94% precision for AI slop. As of August, members saw 40% fewer views on content classified as slop.

Real names are the moat

At its core, LinkedIn is about authenticity, credibility, and trust. All three are what the AI flood is calling into question everywhere online.

LinkedIn's distinct advantage over other social networks: members are required to use their real names. That inherently builds more trust than the pseudonyms popular elsewhere. And it creates a self-policing mechanism. People are less likely to show the worst versions of themselves when they don't want it to hinder opportunities down the line.

The community flags it, the AI learns it

I especially like the approach they're taking.

The community notifies them what they think is AI slop, and their AI systems learn what it is about that content that rubs people the wrong way, so they can identify it more intelligently and proactively, without our input.

Substack tried something similar a few months ago and it sort of backfired. They used a detector to determine if something was human or AI slop, and to what percentage. People ran things through the Pangram AI detector that they wrote before AI was a thing, only to be told they used AI to write it. Not a great look.

The sentence that should worry repetitive posters

The other thing that jumped out at me:

"Now we're evolving our systems further to better recognize and reduce repetitive content."

That is a very important sentence if you use LinkedIn to brand yourself, your company, and generate leads for your business. I've seen countless small business founders complaining that their reach has been decimated. Mine has declined steadily over the years too, and September 22 to October 1 was the lowest performance I've seen on LinkedIn in years.

But what that sentence means to me is that if you're always posting the same thing over and over and it's overly promotional, LinkedIn will limit your distribution on that front alone. The popular Instagram CTA "Comment X to get Y" is also being targeted for extinction by LinkedIn as are engagement pods.

The bottom line is that LinkedIn is actively trying to expunge the reputation that it has become a platform for pitch slaps and growth hacker gurus, since that's not aligned with their mission.

So instead, consider getting more creative and more personal with your posts.

Your perspective is 1/1.
Your personality is 1/1.

You can use AI to put both out there online, but if you use AI without your perspective or personality, you're just another lazy person checking a box, and that's not what LinkedIn wants. I'd argue it's not what the people who matter in your network want either.

LinkedIn is becoming a branding platform.

Employer brands.
Personal brands.
Brand marketing. (This one is lagging behind, in my opinion.)

So you need to ask yourself:

Do you want to be known for having an authentic, reputable brand?
Do you want to connect with humans and build real relationships?
Or do you want to cut corners and put in the least amount of effort?

Anyone in marketing will tell you branding is a long-term strategy. It's top of funnel more so than bottom. If the top is broken, there is no bottom, because LinkedIn simply won't bother showing it to anyone.

Use AI. Do not outsource your thoughts.

So go ahead and use AI to help you craft your perspective and synthesize your thoughts into something perhaps more cohesive than you feel capable of writing.

But do not outsource your thoughts to AI.

Hi5,
LD
🌶️