AI slop on LinkedIn: When does a post become spam?

Creating LinkedIn posts is wonderfully easy. Define your topics, write a prompt and, in next to no time, AI has churned out your posts. Lots of content for very little effort. But increasingly, that can also mean less reach – because LinkedIn is taking steps to curb the practice.

LinkedIn users can report AI-generated posts

Since 30 July, LinkedIn users have been able to flag posts as potential ‘AI slop’. The term refers to low-quality, AI-generated content produced at scale. In practice, users can report it via the three-dot menu next to any post by selecting ‘Seems like AI slop’.

Flagging a post helps train and improve LinkedIn’s systems for detecting AI slop. The post also disappears from your personal feed. If an account receives a large number of reports, the account owner will see a notification in their post analytics. Reports are anonymous, so authors cannot see who has flagged their content as AI slop.

Up to 40% less reach

Being classified as AI slop has consequences. LinkedIn’s Hari Srinivasan has said on the platform that content identified as AI slop can see its reach reduced by as much as 40%. In other words, it is shown to significantly fewer people and receives fewer views.

Since the new feature was introduced, more than one million LinkedIn users have already clicked the button. That suggests there is a sizeable group of users who have had enough of AI slop.

Data analytics company Pangram has also looked at just how widespread AI-generated content is on LinkedIn. Its analysis of more than one million social media posts found that around 30% of shorter posts containing 50–250 words were AI-generated. Among longer posts of more than 250 words, the figure rose to 41%.

Who decides what counts as AI slop?

There is, however, an important distinction: AI-generated content is not necessarily AI slop. And who gets to decide what qualifies as slop in the first place?

An algorithmic AI detector can get it wrong. So can the people clicking ‘Seems like AI slop’. There is no foolproof way of proving that a piece of content was generated by AI. The figures can therefore only provide an indication and should not be considered representative.

There is another, potentially problematic aspect too: the reporting feature introduces a new form of social judgement. Much like leaving a Google review, there is nothing to stop someone from clicking the button simply because they dislike the author or want to harm the account’s performance.

Be specific – and above all, authentic

So what does all this mean for anyone posting on LinkedIn? Of course AI can help with writing. Claude, ChatGPT and Gemini are excellent sparring partners for generating ideas or improving and structuring copy.

But for a post to be genuinely good – relevant and worth reading for its intended audience – it usually needs a distinctive point of view, an original idea or authentic insights. And those generally need to come from people.

So how do you make a post individual enough that it no longer really matters whether AI helped with the writing? By drawing on your own experiences, real situations and specific examples from your day-to-day work.

What did a client say that took you by surprise? What went wrong on a project? What decision did you make, and why? These are the kinds of details that make posts credible, tangible and, above all, authentic.

Because ultimately, we as users want the same thing LinkedIn does: to connect and engage with people on the platform – not machines.