The ChatGPTisation of my writing

• # ai, # writing • 339 words • 2m to read

In this video by Cal Newport alongside one Brad Stulberg (chair-mountain?) he mentions how the class of writers have somewhat acquiesced to AI, and so all have the generic, generified manner of writing: the rules of three, the em-dashes, the not-this-its-this, etc.

I don't use ChatGPT (or Claude, my latest vice cum muse) for writing words, though I do use it to assist in programming. I know that's a problem in its own right and am happy to give up that ground: it's a current moral failing, like my use of plastic, and all whatever else: it's a known, conceded vice. Doesn't make it better.

The issue is with the bleed of everything into everything else. If I even use AI for anything, I'll have that taken into myself: the self is a composite of all things it experiences. Also, if I read writers who use AI, I'll take that into myself as well.

Another point they make is that AI is essentially the car for the brain. It provides massive speed and utility, but at the cost of atrophying the muscle, just as a car atrophies the muscles of locomotion. I think also, not mentioned in the talk, that the metaphor can be expanded, in that the car spreads itself in a way as to make its harms unavoidable (noise and air pollution for instance) just as AI affects the "atmosphere" of knowledge and in so doing makes its deleterious side-effects affecting to all. The damage will also be in altering the knowledge-using or knowledge-creating landscapes in such a way as to be dependent on AI, again cf. cars and motonormativity.

So, there is a strong argument here to take harder-line approach to AI, so as to prevent this affectation. There is also a strong argument to use it like a car: a car does have utility and should not be avoided totally. For certain workloads it is useful. I think the issue is in becoming the kind of person who takes the car to the corner shop.