Collaborating with AI
Or having a computer on your team...
*** Updated with correct link to ChatGPT session in the copy - unfortunately it can’t proof-read that :)
Lately, I (like very many people) have been exploring using AI in my work. Indeed, by the time you come to actually read this, ChatGPT will probably have proofread it for me, which means there will be semi-colons in it, slightly more eloquent use of language, and it will probably be spelled correctly too (thanks, Chat!).
The conclusion I’ve come to is that ChatGPT (and other AIs by extension) aren’t simply tools that I use - they’re team members to collaborate with. Quite often, I’ll:
Get Chat to write me a first draft of something so that I can get some ideas of where to go with it.
Write my own version, taking elements of what it’s produced but adding my own real-world observations.
Give it back to the AI for proofreading.
I also, perhaps weirdly, feel the need to be nice and polite to the AI. I think it’s somehow important to me psychologically;1 otherwise, I might feel like I was taking advantage or something. Do you do that too?
Anyway, here’s an example of such a discussion based on an article I recently published on Kempt & Co on ‘The Danger of Cognitive Bias.’ As you can see, the one produced by Chat is pretty good, but I felt it was a little too general, and the steps to avoid the problems weren’t clear enough. I hope you agree that mine is better. Happy reading!
Links:
Article on Cognitive Bias: https://www.kemptand.co/content/cognitivebiases
Chat GPT session: https://chat.openai.com/share/3cdf8788-2a6e-453a-858b-5c230e4cbd54
Note the semicolon added :D