[I recently posted about my experiments with ChatGPT. I also polled people on how comfortable they were with using AI for writing, from just checking for errors up to creating something for you, in your voice. See the fascinating debate with huge range of responses on LinkedIn.]
I’ve used it mainly to edit. I’ll paste a draft and ask if anything is unclear or missing, or simply say “Tighten this.” And I often think, “Oh, it’s right, those 3 sentences could be one”…and I then edit my own piece (not cut-and-paste).
But yesterday I tried something new: I gave it half a piece and some notes and said, “Finish this.” What came back looked really good and, well, sounded like me. [I didn’t end up using much of it, but still.]
I asked ChatGPT HOW it writes like me. It then accurately described my own style to me in a few bullets. And now it has read every article on my blog (where I re-post most of my published work). It can recreate me.
So here’s the question:
How do we feel about using AI — trained on our own writing — to help create something new?

- If you enjoyed this post, please pass it on. Subscribe to get all of Andrew’s articles in your in-box.
- Follow Andrew on LinkedIn and BluSky
- Please check out my new TED Talk on Courage
- Join the Net Positive movement or to expand your learning, check out my new LinkedIn Learning class, two Net Positive online classes, and other executive education offerings.

2 Responses
Fascinating. I am also using Chat and other AI’s this way. I always check and too frequently find misinformation when I ask it to research something. But I consider it a good junior partner in my work.
“Junior partner” is a good phrase for it. And yes, there are errors. For anything data related, i ask for the link and go check it out.