December 11, 2025
It’s been a while since AI really took off on social networks. On LinkedIn, it’s been a mixed bag – and we don’t see AI as all bad, but we do see the limitations. If you’re doing paid or doing one-click content/responses…we advise you to read on.
Firstly, some stats from OriginalityAI:
- 189% Increase in AI Posts on LinkedIn Since ChatGPT Launched
- As of October 2024, 54% of long-form posts on LinkedIn are estimated to be AI-generated. (probably higher now)
- AI-assisted long-form posts show an increase in word count by 107% since ChatGPT.
The Organic Side: Writing, But Make It Sloppy
LinkedIn’s Premium subscribers now have access to AI-powered writing assistants that can help craft posts, optimise profiles, and even summarize lengthy content. Because nothing says “thought leadership” quite like letting a language model think your thoughts for you…
So, never struggle with writer’s block again, right? Get post suggestions, sound more professional – but to us, when everyone has access to the same AI assistant, trained on the same corpus of LinkedIn’s existing content (already a cesspool of buzzword salad and humble brags), we’re creating a feedback loop of…mediocrity.
AI helps us do things fast, but the idea it’s been used as a ‘co-pilot’ is not the truth. It is being used to craft and create, reduce thinking and therefore nuance in messaging. What’s the differentiator? Well, there isn’t one.
For those thought-leaders who thought AI would help drive their engagement and impressions – congrats, it may be working. But it’s also working for the other guy. And that other guy over there.
Oh the irony
LinkedIn has spent years positioning itself as the platform for authentic professional connection. “Be your authentic self!”, “Share your professional journey!”
Then they hand you an AI ghostwriter and say, “Here, let the machine be authentically you.”
Something of a cognitive dissonance we feel.
We’re supposed to believe that AI-generated posts about “lessons from failure” and “what my team taught me about leadership” represent genuine human experience. We’re expected to engage with content that was literally manufactured to maximize engagement, not to communicate something meaningful.
‘Leave a comment’
Perhaps the most obvious AI farming tool on LinkedIn right now is the auto-comment. Problem is everyone knows it’s an AI comment – it’s almost as impactful as hitting ‘like’. What’s more, it doesn’t actually reward the content. LinkedIn itself says that comments above 5 words and especially over 10 are far more likely to be rewarded, so even using their AI tools in this instance is pushing against a closed door.
The Paid Side: Accelerate Away Your Control
On the advertising front, LinkedIn has blessed us with Accelerate, their AI campaign creation tool that promises to build entire campaigns with minimal input. Just tell it your goal, set a budget, and let the machine do the rest.
Trouble is, less control in this case isn’t good. And if we’re being honest here, we’re relying on LinkedIn to understand your marketing message properly, as well as their own tool (built by a human) to understand marketing well enough in itself.
We have experimented with this tool and…it’s limited. There are many layers to a paid campaign. Picking simple objectives is a starting point, but pulling artwork from a CTA link and creating some very limited visuals is not going to deliver what’s needed. The copy is also very obviously AI. I would also advise LI to look at IP addresses – we don’t all want US spelling…
Sure, Accelerate can help to push through things quickly but proper creative needs a creative human mind and eye. Whether you’re leaning on agency or your creative director, we feel LI is well off being able to make your campaigns really pop.
There is nuance in the flow of animations; speed, graphic fades and word speed. With carousels, design elements are almost more important. Remember that you know your campaign better than anyone, don’t just use tools for the sake of it.
And as for LinkedIn Audience Network…well, that’s for another blog.
AI Ad Variants: Creativity by Committee (Where the Committee Is One Algorithm)
LinkedIn’s AI can now generate multiple ad variants automatically, testing different creative approaches to see what resonates. It’s basically the platform saying, “Don’t worry your pretty little head about creative strategy. We’ll figure out what works.”
And what works, unsurprisingly, is whatever keeps people scrolling on LinkedIn, clicking ads, and generating revenue for… LinkedIn. Not necessarily what advances your actual business objectives, builds your brand identity, or creates memorable campaigns. Just what performs well within the walled garden’s metrics.
We’re not saying it’s bad or wrong all the time, but it is designed to speed up what you do so you can go and do more. Click more, more data, more advertising revenue. Ad-nauseum.
What This Really Means for Marketers
When your content strategy is dictated by what an AI suggests based on what performs well on platform, you’re not doing marketing. You’re doing algorithm appeasement.
We’re losing differentiation. If everyone’s using the same tools to create “standout” content, the only way to actually stand out is to not use them. Which creates a weird arms race toward either AI-generated sameness or deliberately anti-AI positioning.
Thus, we’re ceding strategic control. Tools like Accelerate are positioned as time-savers, but they’re actually removing you from critical decisions about targeting, messaging, and creative direction. You’re trading control for convenience, and the platform is the beneficiary.
How do you learn if not by doing? How do you report to the CMO or CFO if you’re not involved in these key steps to understand why something is performing well or badly?
What’s more, how to do you optimise or fix the problem? AI will tell you ‘spend more’…when that’s not always the solution. In fact, more often than not, it’s a creative or targeting issue.
The Way Forward (If There Is One)
Use these tools tactically, not strategically. Let AI help with drafts, variants, and data synthesis, but don’t let it make decisions. Don’t let it define your voice. Don’t let it replace the actual human thinking that separates good marketing from AI slop.
Remember, we’re not marketing to algorithms. We’re marketing to people. And people are getting pretty good at spotting the difference between content that was crafted with intention and content that was extruded through an AI content machine.
The Bottom Line
We don’t want to say, “LinkedIn’s AI tools are evil”. They aren’t, they’re just the logical conclusion of a platform that wants growth – and that’s business sense. They’re using AI because everyone else is and don’t want to be left behind or seen as ‘old school’.
But we feel they’re not making LinkedIn better for users. They’re making it more efficient for LinkedIn. Users are already flagging that they can spot the AI a mile off, and that is the problem. People don’t like seeing it – they want to be fooled into seeing it.
Do you have any thoughts? Leave us a comment or contact us.