LinkedIn’s AI-Slop Problem

Why everyone sounds the same — and why that may make authentic B2B content more valuable.

I don’t spend much time on LinkedIn these days, if I can help it. But most days I open it out of curiosity or necessity. And lately I can't help but notice - everyone sounds the same.

Not just broadly similar, I’m talking indistinguishable. Especially when it comes to brand or company pages. All the posts have the same rhythm and pacing, and even turns of phrase. It’s all a very neutral tone that manages to sound both enthusiastic and completely devoid of opinion at the same time.

Usually there’s a hook, followed by a slightly contrarian sentence, then a tidy little sequence of insights. Sometimes there’s even some sort of moral at the end tying everything together with a CTA.

Technically, there’s nothing wrong with any of it.  And yet, there’s a hollowness and a sense of unease attached. Now, I won’t coin this phrase, but what we’re seeing is some sort of industrialization of B2B content. Or, less politely, AI slop.

We’re trained to recognize it, or at least question it, when it comes to videos and imagery on instagram and facebook now. If we care enough to do so, perhaps we should apply the same scrutiny to B2B content?

The irony is, of course, is that the posts trying to stand out end up sounding remarkably similar. While a ChatGPT prompt may have a contrarian idea or even attempt to weave in personality - with the underlying cadence always the same, everyone is always calm, balanced, and remarkably earnest.

Which makes sense. Most AI writing tools have been trained to produce something safe, readable, and broadly agreeable. They’re optimised for coherence and clarity after all. They are  certainly not optimised for opinion  or originality. 

The result is a strange kind of equilibrium where everyone appears to be saying different things but somehow still ends up sounding the same.

The new authenticity signal: typos

One of my favorite responses to this has been watching founders deliberately introduce small imperfections into their posts. Weird  capitalization, the occasional typo, punctation slightly off. A sentence that clearly hasn’t been polished to within an inch of its life.

The message they’re trying to send is simple:This came from me. Sometimes they even feel they have to write that.

I kind of love that writing on the internet has always been associated with polish and precision, yet here we are and the very signal of authenticity is the opposite. When your readers start looking for imperfections as evidence that a human was involved, it means the baseline has shifted.

The feedback loop problem

There’s another, more subtle risk with the current flood of AI-generated content. AI models are trained on existing writing. 

Then  we use those models to generate new writing based on that material. Which then becomes part of the broader body of content on the internet. Which future models will inevitably learn from. Over time this creates a rather dystopian, and slightly existential, feedback loop.

The models get better at producing writing that resembles the average of everything they’ve already seen. Hence you end up with an enormous volume of content that is technically competent and conceptually very similar. Which is exactly what LinkedIn currently feels like, already!

The future of B2B content

The uncomfortable truth is that most B2B marketing problems aren’t really content problems anyway, they’re differentiation problems.

The job  has never been to simply to publish more and more material (and hope some of it sticks, or should I say converts).  Instead, the value is in explaining why something is different, why a particular approach works, why a product or service exists in the first place  all in a way that  matters to your audience.

Those answers require a point of view. And that  usually comes from proximity to real situations and lived experience. That kind of thinking is hard to outsource.

As i write this I realize that this almost sounds like an advert for Claude. Anthropic’s “Keep Thinking” campaign speaks directly to this idea — positioning the tool as a thinking partner for meaningful problems rather than just another AI assistant. (ChatGPT). Do you think Claude wrote that copy? We may never know.

Ironically, the explosion of AI-generated writing may end up making authentic writing more valuable rather than less.

When every argument has been smoothed into something universally agreeable, the writing that holds attention will be the writing that actually takes a position.

And when the internet is full of content  from existing patterns, the rare thing becomes something that more clearly came from experience.

Of course, as marketers, it is totally viable for you to train an AI assistant on a company’s  tone of voice and brand guidelines. If you’re not leveraging AI tools right now, I would be worried about you - we’re all learning as we go. But with every piece of content that leverages AI, you still need a) someone to have the experience and ideas B) someone to write the prompt, and well, and C) if we’re being honest for today, have the skillset to fix the AI output on the other side for readability, authenticity and connection.

AI will undoubtedly make all areas of sales and marketing process faster. It should absolutely be part of the modern content generator's arsenal. But speed was never the difficult part. Thinking still is.