My AI style/grammar checkers consistently disagree on certain things.
For example, they always want to turn “your own whatever” into “your whatever.”
When I use the phrase, ‘It’s your own choice,’ I’m conveying a unique meaning that’s not captured by the simpler, ‘It’s your choice.’
(I do worry that some automated system will try to be helpful and make both quotes identical by removing the word “own.”)
Similarly, when I say, “It actually matters” versus “It really matters,” the two express very different thoughts. (We will not mention what would be lost if it capriciously removed the word “very” from that comparison.)
Each conveys something very different from “It matters,” which the systems want to reduce both to. (And I don’t think unsplitting my infinitive would help here.)
Most of the time, it wants to remove words that it thinks are non-essential or “make me sound uncertain.”
And, while I am not quite sure, perhaps, maybe, it might actually be that I could, at some particular time, indeed, truly be uncertain, don’t you think?
And while I do endorse brevity (which has fewer letters than conciseness), I sometimes wonder whether it is optimized for minimizing my telegraph bill, where I would pay by the letter.
Mainly I feel like it is trying to remove my “voice” from my writing.
Perhaps that is because voice varies over its training set, so voice gets averaged away (?)
Yes, I should fine-tune a model to learn my voice, and perhaps I will at some point.
But today, my thoughts are not on how to fix this for myself but rather on what happens when a generation of aspiring writers grows up with these tools critiquing their writing before they develop their own voice.
I fear that it is not anything good.
