ChatGPT wrote my LinkedIn posts for a week. Here's what happened.

For the last week, all my LinkedIn posts have been written not by me, but by the ChatGPT AI. 

All the accompanying pictures have been generated by ChatGPT’s sibling, Dall.E. 

Both of these in response to text prompts I gave the machine. None of the copy posted was written by me. It was all generated automatically.

In response to my request “Write a LinkedIn post about how to create YouTube thumbnails which maximise click-through rate”, chatGPT 'wrote' this:

The image was artificially generated when I asked for “A photograph of a woman looking surprised and happy, in close up”.

LinkedIn users responded positively to ChatGPT posts

The posts created have generated 2,500 views and 24 engagements. A handful of you have responded positively with likes or comments. I’ve even had a couple of interesting real-world conversations off the back of them ( 👋 John Peddie ).

I experimented with a few different topics: I published three posts about YouTube, one about teamwork and another about leadership during difficult times.

Interestingly, considering they were written by machine, the post with the most emotional content generated the most engagement - read that one here:

The image there was generated in response to the text prompt "a man helping another man as they walk through a dark forest, in an impressionist style"

The posts have been, largely, reasonably well written. They’re well formatted and make sense, and do provide *some* value. ChatGPT correctly tailored the content to the platform: it generated copy of an appropriate length and in a tone of voice common to LinkedIn. It even added hashtags!

BUT

It seems clear to me that the posts are not in my tone of voice. The AI uses words and phrases I wouldn’t, including an over-use of exclamation marks!

Although there is some value to the posts (I thought the one about YouTube thumbnails was actually pretty close to the advice I would have offered to business starting out) it was generic and - I’ll labour this point - derivative.

What I learned

The output from chatGPT is derivative by nature. All the AI can do at this stage is amalgamate information from the copy or pictures it’s been trained on. So it can do a convincing impression of a LinkedIn post, built from examples of previous LinkedIn posts it has ‘read’, but there is an important element missing: the value of real-world human experience.

The most useful LinkedIn posts I read are those written by practitioners I admire. Anyone can read and regurgitate content. But not everyone has the opportunity or the skill to practice their craft at the cutting edge and distil their real-world learnings into actionable insights for others.

This is something that, right now, the machine can’t do. And that’s why I won’t be using AI to write any more LinkedIn posts.

Instead I’ll focus on sharing my real world experiences with other humans (you) with the aim of all learning together.

 #AI #LinkedIn #writing