The ChatGPT experiment on LinkedIn is almost over, with tomorrow marking the last day of AI-generated posts and images. Looking back, I’ve realized that writing posts via AI is not all that different from writing them completely by myself. However, I did find that the more concrete I was in creating the posts, the better their content turned out.
Answering others in the comments was a bit more challenging, as the results never really mirrored my personal style in writing LinkedIn comments. On the other hand, creating accompanying images was a blast and something I will definitely continue doing in 2023.
While I do believe that text AI can be helpful in creating topics and themes for posts on LinkedIn, I don’t think it’s appropriate for creating the actual post itself. I noticed a pattern where the AI would list and explain various main concepts or topics in a post, and then finish by summarizing everything and refining a main message. However, this isn’t always something we may want in a post.
I received mostly private messages as reactions to my posts, and even answering those via AI made the discourse feel a little „robotic“ at times. In the future, I will definitely not go down that road when it comes to building up my personality on social media.
Overall, I’ve learned that while AI may be useful in helping with the topics and structure of a post, it’s not appropriate for building up an authentic persona.
PS: I even used AI to create this year’s Christmas cards, including both the card image and text with the help of Stable Diffusion and ChatGPT. If I hadn’t told my friends and family, they probably wouldn’t have noticed.