LinkedIn changed what it thinks is important: Here’s what you need to know (October)
LinkedIn changed what it thinks is important: Here’s what you need to know (October)
LinkedIn is getting better at identifying AI Slop and it wants to stop the spread
It’s no secret that the social media platforms have tasked themselves to fight back against the rising volumes of AI Slop, and now LinkedIn is evolving its approach once again.
In a post entitled “The Technology Behind LinkedIn’s Fight Against AI Slop”, VP of Engineering Tim Jurka offered some insight into the technical advancements made in the crusade:
- LinkedIn is refining its AI systems to better identify AI Slop
- User feedback is helping to train the systems themselves
- The system is getting better
- It has expanded the remit of the system to prevent contagious spread
The approach is working – the systems have 94% precision in identifying AI Slop and have reduced the views of AI Slop by 40% thus far
The important point to note here is that the AI systems being used are not just identifying AI Slop, they are actively limiting its distribution.
Identifying the Slop
Back in June, LinkedIn said:
“When content appears to be generated by AI and lacks clear perspective, it is less likely to be widely distributed beyond a person’s immediate network.”
That is a bold statement of intent, and his latest statements this month confirm the team is following up with action.
What is important to understand is that this is targeting poor-quality content generated by AI. The systems are trained to identify a number of different indicators, such as:
- Vague and generic language
- No distinctive opinion or perspective
- Repetitive or recycled content
- Content to game attention, effectively click-bait content
- Low value, automated comments
LinkedIn hasn’t explicitly disclosed the precise technical indicators, but it has described the characteristics of poor quality content. The above is us reading between the lines.
This is not to say that AI generated content could not be good quality, but it takes skill and insight to craft good quality AI generated content. LinkedIn’s interest is to protect the user experience, this is something every marketer should remember. It is actively seeking to limit the impact of poor quality content and promote good quality.
Halting the spread of the Slop
The most pressing concern for marketers would be the consequence of the Slop. A post which is flagged as AI Slop will be visible to the direct profile of the author, but perhaps not all of the connections, but distribution beyond those connections is less likely.
The value of social media platforms as a distribution vehicle is to engage known and unknown audiences. A post can gain traction beyond immediate audiences, and with some being actively promoted by the platform to those users it deems might have an interest.
But that spread is restricted when AI Slop is detected, radically minimising any potential gains.
Once again, this indicates the social media platforms are not fans of short cuts when it comes to content, which has a direct impact on how useful their platforms are perceived to be. It is within their interest to push back against poor-quality content.
The inferred lessons for marketers are therefore as follows:
- The volume game is dangerous – identify good ideas and focus on those alone
- Original thinking matters – speak to people within the organisation to identify ideas that resonate with your audience rather than relying on AI to come up with something polished yet generic
- Give executives a voice – business leaders can humanise the brand, while LinkedIn has said it wants authentic perspectives and lived expertise
- Think audience first, not company message – start with what people want to understand, rather than what your business wants to tell them. Focusing on the company narrative alone risks content becoming repetitive and generic
If content is to remain king, it must be built by those who know how to create impact and elevate ideas into something
