There’s a common refrain, among academics in particular, that AI-generated writing is ‘sterile and neat‘ and that it lacks (for want of a better word) ‘humanity’.
The position seems to be that, once you’ve handed off a piece of writing to AI, you’ve effectively lost the option of expressing any part of yourself in the output.
But it’s worth noting the following:
- When AI generates output that’s sterile, it might be because that’s what a lot of human writing looks like.
- Naive prompts lead to poor outputs. Better prompts are more likely to get you closer to the results you’re looking for.
- Recognise that you’re in a writing partnership with AI; if the output is weak, it’s partly your fault.
- The first AI-generated response typically isn’t the best response. Use that first output as the starting point for an interaction with the model.
If you think the AI-generated output is sterile, prompt it to write something more messy and less neat. You can just ask it to write something you think of as more human.
And once it’s written a draft you’re OK with, there’s nothing stopping you from going in and adding your personality. I don’t see why it’s almost always positioned as all-or-nothing with AI. You can work with it to create something that’s both part of it, and part of you.
AI-supported writing can be whatever you want it to be.