I should be sleeping

[dhampyresa]

I'll confess that when I distantly saw people talking AI and racism on AO3, I thought someone had suggested using AI for content moderation and I went full gun_by_my_printer.jpg



But those are unrelated things! Thank fuck.

Also I continue to not have heard any pro-AI argument from anyone who knows what they're talking about on both the creativity and tech side. Although I have to say I don't think I've so far seen anything as egregious as that guy who wrote 40+ tweets that boiled down to "with AI anyone can write a novel" -- which. If you can produce this many consecutive tweets? You always could, you just never did.

At present my opinions on AI are: (1) it shouldn't be called that because that makes the steroided autocomplete that are large language models (and visual art equivalent) sound like JARVIS or HAL to people, (2) there are major ethical issues with the datasets as they currently exist and are used and (3) even in the very improbable case that all copys of all versions LLMs are retrained on ethically sourced data, I'm not interested in the output, either as "help" for my own creation or as something to interact. Also why does no one ever share what prompts they use?

The racism conversation is in reference to endotwracism which as far as I can tell is a well-thought out initiative?

I think I'll try to post the Shuri vs the gods of pre-dynastic Egypt fic for it. I need to finish writing explanatory notes for it -- and find a title. It'll probably be something from the Book of Coming Forth by Day (because significant parts happen in the Duat) or from Épitaphe by Léopold Sédar Senghor (because the parallel between its ending lines "Quand je serai mort ma Signare, couche moi sous Joal-l’Ombreuse / A l’ombre des Ancêtres" -- "When I am dead, my Signare, lay me down under Joal-l’Ombreuse / in the shadow of the Ancestors" -- and Killmonger's "Just bury me in the ocean with my ancestors who jumped from ships, 'cause they knew death was better than bondage." was one of the sparks behind its inception). Then again I wanted to post before Moonknight came out. Then before Black Panther 2 came out. So who knows. Presumably I will remain a slug.


In conclusion... CPR in space is done via handstand? (Source on PubMed: Cardiopulmonary resuscitation (CPR) during spaceflight "We recommend a differentiated approach to CPR in microgravity with a division into basic life support (BLS) and advanced life support (ALS) similar to the Earth-based guidelines. In immediate BLS, the chest compression method of choice is the Evetts-Russomano method (ER), whereas in an ALS scenario, with the patient being restrained on the Crew Medical Restraint System, the handstand method (HS) should be applied.") Yeah, I got nothing.

Comments

[muccamukk]

Oh, good, the nipplebot, but for something even MORE difficult to pin down than what a dick looks like. Does no one remember what happens when we automate content moderation!?

ETA: To be clear, I thought that initially as well, and was impressed that someone had come up with something *worse* than user generated tagging.

[kore]

I gotta admit part of me thinks it would be meanly HILARIOUS to let loose a moderation bot on AO3 but then again I'm a Scorpio.

It sounds like AO3 volunteers are terminally overwhelmed just trying to keep the site going as it is and they're saying that's one big reason change re racism is not happening. I dunno. It seems like a volunteer workforce is not the right approach for what it's become, but that's also such a huge part of the ethos? Idk.

It sounds like a lot of ppl are way less happy with the "we're overwhelmed, we're having committees about it, we're radically inclusive, we can't grandfather millions of fics in to a new category" excuses when it comes to botfic. As opposed to racism.

[muccamukk]

Yeah, I 100% believed that they haven't banned computer generated fanworks because they can't verify bot v human with any accuracy. Still sucks.

[kore]

Yeah, it's not AI! It's not an intelligence! I think if ppl realized it was a lot more like "check all the squares containing a bus" or citizens training facial recognition for law enforcement, fandom would be a lot less entranced. But I've seen a lot of pol playing around with it like it's a prompt generator, at best, and going "I wrote an AI fic for this fandom with three fics in it!!" and let's not even go there.

Re otw legal, all I can think is either they're trying to defend botfic because they think arguments against it will be used against transformative work, or....idefk. fair use maybe -- a lot of ppl are arguing botfic is plagiarism, but it's not, altho it's not original, either.

"I think I'll try to post the Shuri vs the gods of pre-dynastic Egypt fic for it"

Oooh, lovely.

[hokuton_punch]

Shuri vs the gods of pre-dynastic Egypt!! *__________* I hope you do manage to pick a title and finish the notes because EEEEEEEEEEEEEE.

[flamingsword]

I do know some disabled folks who use AI image generators as assistive technology, so that's one plus in the pro-AI column.

Also, I don't really know that we know how human visual intelligence works well enough to say whether or not it's at all similar in nature to image-generation AI. We don't have artificial general intelligence yet, certainly, and that's probably a good thing. But the fact that we know how to create even a limited LLM or image generator is already the stuff of sci-fi, and we don't know enough about ourselves or have a unified theory of what a mind is to be making any hard rules about it yet. And unfortunately, this is one of those fields where we're going to have to run the experiments first and back-figure the rules later.

I don't think banning AI content is going to be much of a help to the artists the AI was trained on, either, since you can copyright images but not styles. As long as the image is not an exact reproduction, that's fair use just like it would be for a human person who made tribute art in any given style, a la the Andy Warhol soup cans.