This essay showed up linked on Metafilter, and I perhaps will note: maybe don't read the comments (someone comes in and basically suggests running it through an AI to "de messify it" and to "stop making everyone a stereotype" and then people start fighting back and forth. There are few things on MeFi that DON'T lead to a comments-fight).
But I found it interesting, even if it's a different style than I'd write in (but that's also part of the point of my post). The writer here describes himself as Blind (he also describes himself as gay and as "an obscure writer")
He describes - I have no idea how fictionalized this account is, but it feels true - how his writers group devolves when a couple AI bros move in and push people to use things like Claude to generate their prose
The Colonization of Confidence
In particular, I was struck by what happened to Leo - who wrote powerful and surprising images (collard greens' smell "subduing the air in the kitchen," which, yes, anyone who's been somewhere with some kind of mustard family plant cooking nods in recognition, even if they never thought of it that way before) that get smoothed out to a bland pabulum when he starts using an AI to "help."
And he gradually loses his voice - because he loses his confidence. Because he believes AI is better at it than he is, culminating with him successfully selling an AI written essay to a magazine. (I recoiled in horror at that bit).
And the writer of the piece describes making a NEW writer's group, one that doesn't use AI, one that embraces the messiness and humanity.
And that's what struck me. That's part of why I dislike AI taking over "art" and "humanities" in particular.
When I read, I want to be surprised. I want to be made to think in a way I'd not thought before. I want to see someone else's experience that is different to my own, and, sitting with it, try to understand it. I don't want some kind of bastardized average human experience, smoothed and "corrected" via an algorithm that is trying to make the most "marketable" thing.
I don't want writing that is like a slick tv ad! I want the messiness and humanity and genuine emotion!
And I don't know how we stop the onslaught of AI, other than, maybe, refusing to buy books written with it - which may eventually become "buy no books written after 2022" because just as bad money can drive out good, I suspect stuff like AI may drive out the quirky, real, genuine-ness of writing.
***
I used to write. I was never very good. I wrote poetry, a lot of it (I posted a little on here years back). Most of it, as I said, wasn't good, which was why I never submitted poetry or stories anywhere. But it was enough for me to have written it. It was enough to take the idea I had and either scribble it down in a bound blank book, or type it into a Word file, and edit it a little, and then periodically look at it again and think "I wrote this. It came out of my brain, and it helped me deal with the world."
I wish more people would just do the messy writing and put it out there.
I wish there weren't so many people looking to profit by promoting what some call "slop engines" - the algorithms that turn everything into something inoffensive and the equivalent of a simple bland frozen dinner - unchallenging to the palate and maybe actually not that nutritious and probably stuffed with artificial ingredients. I like the messiness and the reality.
And also: sometimes there are - and there SHOULD be - things you just for yourself. Like the poems I never tried to publish or share. Sometimes I wrote them because an image popped into my mind and I wanted to play with it and develop it; other times, I was having Big Feelings and sitting down and writing about them helped me process them. I am not at all convinced telling an AI "I am in deep grief, write a poem about someone grieving" would help in that way.
AI writing doesn't have the element or surprise or whimsy or novelty. I admit sometimes when I post a riposte to a joke someone made on Bluesky, or make a standalone joke myself I think "Hah! an AI couldn't do that." I am good at "lateral thinking" when my brain is having a good day and I'm firing on all cylinders and it feels to me like an AI can't do that kind of leap-of-thought that is a very human thing, and that some humans enjoy both reading and doing - the whole "Aha!" moment thing.
And I admit, I worry about it. I worry about what happened to Leo happening to a lot of people, I have had fellow profs (mostly at other universities) report that their students don't "trust" their own writing any more and feel like they have to use Chat GPT or Claud as a "crutch" to help them write. Or that it's "easier" than doing the hard work of generating a draft and then editing it (And an aside: I yell at my students about how you really should probably spend at least twice as long editing a paper as you spent writing the original draft. An entire YEAR of my Ph.D. program was just me doing rewrite after rewrite of my dissertation, and finding additional sources or cutting out some references or doing the analysis a little differently and rewriting with that)
I don't know if I'm being alarmist when I think "we are going to lose something of what makes us human if we give too much of this kind of thing over to AI" but other days I do not think I am being alarmist enough about it....
No comments:
Post a Comment