More than half of the long-form thought leadership on LinkedIn is generated by AI, according to recent research by Jonathan Gillham, founder of the startup Originality.ai
Gillham said the introduction of ChatGPT has caused an increase in the number of longer posts and a surge in the number being written with AI.
The study analyzed 8,795 LinkedIn long-form posts (defined as a piece with more than 100 words) over a period of 82 months from January 2018 to October 2024.
Gillham found that in October 2024 about 54% of those posts were likely generated by AI. The average post length has more than doubled since ChatGPT was introduced two years ago.
As a ghostwriter for CEOs who publish primarily on LinkedIn, I don’t recommend using AI because people can feel the difference and research shows readership and engagement is lower on machine written content.
It’s like sending an electronic e-card for a birthday instead of Hallmark.
You are sacrificing effectiveness for efficiency.
You can use a machine to help, but you cannot be the machine.
(Note: I do recommend using Perplexity, ChatGPT, Claude, Gemini and other LLMs to do research and check if subjects have already been written about.)
If you are writing on online with the aim of being a thought leader you probably already know that only 1% of people on LinkedIn post. If you are going to the effort, don’t reduce your engagement by using AI.
I ran one of my recent posts about Union Square Ventures through the Originaltiy.ai calculator. It had an overall score of “Likely Original” with a 100% confidence score.
However, the application also rates each sentence and the last three lines were scored as “likely AI generated” at thresholds of 31%, 44% and 46%, respectively.
Here’s the last sentence: “Increasingly, the only obstacle will be the willingness of a company to ask hard questions.”
I wrote that myself, but I’ll acknowledge it wasn’t my best work. It feels and sounds canned.
It’s a reminder that generic content doesn’t work, whether written by computer or human.