My Artificial Accomplice

This is a post about utilizing AI to help with writing… which I’m writing without AI.

I’m not opposed to using it. But I’ve had reservations. For the same reasons I’ve hesitated to hop onto some social media platforms and why I postpone getting my middle-school kid her own cellphone: I’m concerned about the dehumanizing effects of these innovations.

Don’t get me wrong. I’m not some luddite who can’t separate reality from science fiction. I know it’s a common plotline these days to have smart devices become self-aware and destructive; before that it was genetic modifications or nuclear radiation creating mutant monsters. These cautionary tales about handling new tech responsibly have been around since we first discovered fire.

So how do I harness ChatGPT’s power without letting it control me? I’m still learning, but I have found it to be an effective writing partner.

I’ve collaborated throughout my career. Writing screenplays with someone else kept us both motivated – like a workout partner reminding you not to skip leg day – and prevented the loneliness of staring at a blank space. But the best part was when we bounced ideas off each other to steadily improve the script. One person’s thoughts would inspire the other to brainstorm new takes which we would build upon and fine-tune until we created something half-decent that neither of us would have come up with on our own.

I pivoted somewhat naturally from writing feature specs to working in unscripted TV, because now I was curating input from a team of writers, producers, editors, and (for better or worse) network executives. And throughout all these experiences, I’ve learned to handle the egos and personalities (my own included) that comes with swimming in a think tank.

Now, dabbling with AI, I’m finding that I can get most of the benefits of a group effort without anyone’s feelings getting hurt. When I present ChatGPT with some rambling concepts, no matter how chaotic they are, it’ll pitch ways to organize and expand on them, following the “yes, and” philosophy and never shooting anything down. And if I pass on its suggestions, it won’t argue or grumble about not being heard.

But interestingly, I still try to show the same respect to my artificial accomplice as my human ones. My philosophy is that you can’t nix an idea without offering an alternative one. And though I don’t have to validate its input, I still carefully choose prompts to steer it in the right direction and even empower it to course-correct itself… and hopefully teach it not to over use those double-dashes or format its sentences so formulaically.

I know I’ve only scratched the surface on using AI, and I’m trusting it to help me more and more. So why didn’t I use it to write this post?

Call me crazy, I don’t want ChatGPT knowing I’m talking about it.