The debate about generative AI

[VibrantFox]

@DokiDokiTsuna pretty much covered all i can think of tbh

---

Like at it's first appearance when it was nothing but indistinguishable blobs I figured it would be a short lived thing only enjoyed for its novelty and then everyone would move on. Then datasets started being built of of existing copyrighted works and everything snowballed from there becoming the disaster it is now. Major corporations have either hopped on the bandwagon to profit and the few who actually have some moral backbone have tried to fight it. American legislation has been near useless in the fight since some politicians are too old to understand or they only see the potential profits for them and honestly the only laws that seems to even respect artists in terms of copyrights is the EU (namely referring to a thread that spurred some reading on Europe's ethical copyright which I wish we could have something like that here)

But the biggest thing to me is everything people swear we can't do without AI now was able to be done without issue perfectly fine prior to it's existence. It's both been admitted and observed the prevalence of what's basically learned helplessness growing in massive amounts because people are training themselves to become dependent on AI and making it harder for them to do incredibly basic things. So even outside of genAI I find it to be kind of useless and pointless in existence. And yes i do acknowledge it's presence in other fields but that's medical mapping and quantitative math not "making pretty picture or writing coz we're too lazy to do it ourselves". It makes me furious that people don't want to make powerpoints the most easy bs-able thing on the planet coz all you have to do is plug in information and make it look pretty and maybe smack some pictures or some music on it to seems like you're really doing something but just basic easy-peasy things.

The disability argument was violently and beautifully debunked by disabled individuals, the accessibility on cost argument was wonderfully dis-proven by creatives in a ton of fields and by ai shills themselves when they showed how much they were spending on their ai bullshit. So there's really no argument in it's favor that I can buy into because I can do that already or I know where to go to find the thing I'm looking for or i can frankenstien something together myself a la old school photobashing or something similar and there's no theft or environmental harm to worry about

Like specifically on the topic of environmental damage alone i do not want a single thing to do with ai because it's damn near reversed so much stuff we finally fixed over DECADES of work. Powering it is killing the planet. The pollution it creates is not only harming the planet but people directly (look up the bbc video on colossus pumping out methane into the air of a city in Memphis and get back to me) and it's making EVERYTHING more expensive from computer parts to electricity bills to water and so on.

Like if AI has a million haters I'm one of them if AI only has one hater left it's me. because everything about it seems so pointless and wasteful and lazy that I just do not care to see or hear about it which is why SO many people do not care to have it shoved in their face and down their throat because we genuinely don't need it people are just to lazy or to scared to try fail and then try again.

Technology was meant to do the difficult labor for us so that we have more time to create but instead it's doing the tasks people deem frivolous so that we can keep slaving away for pennies and thank the folks giving us scraps for being generous and I refuse to play that game if i can help it.

I do think it adds some value to human made creations because of the disgusting prevalence of genai slop but as has been mentioned it really only is seen in that way by those who actually know the difference and care otherwise it's just another shiny piece of content to consume before moving onto the next and that's the same mindset that an unfortunately large and vocal group of people have. My only hope at this point is that it changes before we're all choking on air because it's so filled with filth and we're scavenging for clean water or food coz we've torn into every natural resource to fuel that crap.

But yeah sure the shiny computer slop is great or something :]

[Ordinaryaverageguy]

I also think photography is an inaccurate analogy. Both produce images, but for entirely different reasons & uses. Photographs could never look like IRL painting. They were different kinds of imagery. Paintings are really 3D products. Paintings can represent reality in unreal ways. You can probably think of other distinctions.

However now, when most imagery is digital, there's far less distinction between the products. Digital photos can look like digital paintings & vice versa. They're all 2D products.

I fear it may be whistling past the graveyard to make the comparison & feel relieved by it.

OTOH, AI may turn out to be more disappointing in the long run than its creators & fans predict. Time will tell.

[emergencychopsticks]

i don't think monetization is the crux of many people's issues with AI. i absolutely find it disingenuous to try and profit off of something created by an image generator, especially if one were to claim it wasn't, but my personal grievances with it are almost wholly issues of ethics.

there's no world in which large language models and image generators can have ethically sourced data banks without outright stealing. the models only function to the level they can today with access to millions of arts from millions of artists, and obviously most if not all of these artists are not consenting to or receiving proper compensation for their stolen work.

on top of that, the energy and water that the data centers eats up is astronomical. i don't think i need to explain why they use up so much water, and i'm definitely not the most qualified to do so, but the easiest way to explain it to those who don't know is that every llm is maintained by physical hardware somewhere, and that hardware gets hot. one data center can consume up to 5 million gallons of water a day- similar to what a town with a population of 50,000 people would go through in the same time. (https://www.lincolninst.edu/publications/land-lines-magazine/articles/land-water-impacts-data-centers/ , https://www.eesi.org/articles/view/data-centers-and-water-consumption ) those who live near these centers are often without clean water (https://www.bbc.com/news/articles/cy8gy7lv448o ), put in abysmal living conditions for what? a bit of fun?

the current view of ai image generation is very... novel. it's easy to look at as a fun gimmick, and i'd be lying if i said i wasn't tempted to feed my own art to a llm, just to see what it would look like. i also see a lot of artists refer to it as a crutch, or a time- saver and i personally think that's bullshit, respectfully. you'll never get better at a craft if you don't... do it. that's all i have to say to that point- i know that's not really relevant to the specific comment i'm responding to, but i digress.

at the end of the day, i firmly believe that ai- llms and image generation- have no place in the arts. i think ai could have a place in the world, notably in accessibility: live- time dictation for the blind, live captioning for the deaf, etc. the images you provided can most certainly be created better by a real artist- who the hell do you think the models learned from? there's plenty of artists on here, twitter, vgen, etc. who can create that for you. don't have the money to commission them? that real artist is you, then. take the time to study and learn the style you want. i've made a helpful flowchart here:

[jwabeasley]

Did you really have ChatGPT write this meditation on AI? That's how it reads to me. Maybe it's "translated" but either way that's hilarious.

[kyupol]

Opposition to AI seems to transcend party lines... (I understand most of you on this forum are on the left )

https://www.politico.com/news/2025/12/27/we-have-to-reject-that-with-every-fiber-of-our-being-desantis-emerges-as-a-chief-ai-skeptic-00704333

[NOOR_ASAKZ05]

No, I didn't use chatGPT AI in writing it. I researched it myself, but I think using a translator made the language a bit awkward... (English is not my native language).

[Ordinaryaverageguy]

i firmly believe that ai- llms and image generation- have no place in the arts.

I agree. But I think most people may not be thinking about art the way we probably are. Art is not just a product, right? It's also about the methods & the artist's efforts.

Going back to the photography analogy, anyone can snap a highly detailed photo of a scene but not just anyone can capture that scene in oil or pastel, etc. I think one can take artistic photos but the result is not art in the same sense as a painter laboring to capture the image and the feeling of the same scene.

One is just a perfectly flat image. In the other, you can examine the artist's technique & perhaps even sense their feelings about the image. Get up close & see how a few teeny-tiny brush strokes evoked the little birds in the branches. See how they saw fine details in life & showed them to you with only a little bit of paint? See over there where they over-painted a tree to get it right? And wonder how long it took to reach that skill and how long they worked on that one image. Layers upon layers of things to think about.

Sure, digital art is different animal than what I babbled about just now. But there's still a human being in between the picture they saw (real or only in their imagination) and you. A real person's time & effort & thought & feelings went into the work. Even their feelings about how you'd feel when you finally saw their creation.

Last thing, in this now over-long babbling; we always talk about "a work of art", don't we? Because real art has always required work from someone. In a sense, labor brings forth the brainchild of the artist. In AI slop, there is no real work. So, if I may opine: without work there is no art or artist.

[Ordinaryaverageguy]

I write rather fanciful comments to YT videos that I really like and have also been queried about using AI instead of writing them. Are we getting to a place where anything clever or amusing or thoughtful must be suspected of being AI-generated "because no one actually writes like that"?? Damn, I hope not.

[VibrantFox]

my only gripe on this lament is that theres definitely levels of labor to taking good photos beyond just having a good camera or pointing and shooting and that's moslty said as someone who's had photography as a perioheral hobby and passion besides art which is why even ai photos give me the ick

there's obviously levels of work that can be seen in illustrative work with brushstrokes and such, be it digital or traditional work, spotting bits of a sketch layer or how someone carries a line when they draw, but there's equal nuance in photos if you know where and how to look

it's not something i can get into now but it is something id wanted to pour over in another thread focusing more on how to find and use photos of different subjects but again another place and time

[ButterflyEmpress]

I think the car analogy fits better though a lil flawed.

In the US at least: Public transportation sucks. There's no safe place to ride a bike. Some places have no sidewalks. Businesses are spread so far away with roads and carparks taking up the space in between. All because the US is centered around the car. Sure it has it's conveniences, but it's outweighed by the cost of maintenance and cost to the environment. And those conveniences are stemmed from the problems it created. People are forced into car ownership even if they never wanted one. Places that are independent from cars are a minority.

With AI we're losing creative jobs, data centers are a drain, and it's everywhere now. Can't even use a computer, cellphone, or website without AI crap on it. Sure creative jobs will still be around and people still have the urge to create, but they'll be deemed worthless and unnecessary; unable to compete with those pumping out genAI that gets better by the day.

I'm fuming at the fact the SFO Museum had an Afrofuturism exhibit with all AI art. How can they promote black people as the future and not have any involved?!

[JWZagst]

All fair points. The images I shared were formed by feeding screenshots of my Sims to ChatGPT, but you're right, I could have just drawn them myself. But that takes time and I just wanted a few sample images to help visualize them as different styles of anime characters in my mind. Eventually I will be drawing my own webcomic with them, but I'm still practicing and fine-tuning my own artistic style.

The point of my post was that I think it's fine to use AI for reference material so long as the actual monetized work is human-made. That being said, you do make a good point about the energy and water cost. The water part could be solved with a proper recycling system, though. The water used in these data centers doesn't actually "vanish," since no matter can be created or destroyed, it can only change from one form to another. It evaporates, but that water vapor can be captured and condensed back into liquid water. Though the architecture to do that or to create a closed-loop liquid cooling solution would be more expensive, and we know many of these corporations are greedy SOBs LOL.

So I guess, in conclusion, I agree with the things you have said. People should probably either do the art themselves, hire artists for reference work even when no monetization for that reference work is intended, or abandon the projects completely. But it can also be a slippery slope because that's not far off from condemning fanfiction, fanart, drawing inspiration from existing works, etc. Artists don't consent to their work being used for AI, but they also don't consent to fanart being made in their style (unless they expressly support it).

[VibrantFox]

I just wanted a few sample images to help visualize them as different styles of anime characters in my mind.

To me tho the solution to this is simply explore different animes from different eras or production companies even manga or whatever other reference material to see what styles you like then in what free time you (or any other person) have try to replicate them whether that's personal trace overs or full on studies and that's a well known pre-existing method of self teaching. The same goes for the styles of other artists, observe, study, practice, repeat and then adapt the traits you like into your own work.

I think where people kind of fall apart and maybe it's because it's not taught or explained as much or people just aren't bothering to look because "it takes too much time" or "it's too much work' but the best style studies don't just look at overall style but see how different aspects of a piece are put together into a final whole and i'd say that because the general public doesn't really understand or appreciate the process that's why they struggle to measure the value of human art against ai.

but getting back to my point on practicing styles it's the picking apart of minute details and anyone who's gotten real good at spotting ai tells knows that the machines utterly fail at those so anyone trying to replicate from that will also fail. it's why i get sad that people keep turning to ai because resources exist that don't rely on it all they gotta do is ask or look and yeah i know it feels impossibly hard to find it these days but they are out there.

like when i was looking to find i think it was color/lighting reference i remembered a site i saved that has movies saved frame by frame so you can look through em. movies, tv, cartoons, anime, a pretty good chunk of different kinds of media and movies both new and old. so i went through those and tried to find ones that had the more dramatic lighting i wanted to experiment with portraying in my own art. saved a bunch i liked. started doing paintovers and rough sketches, then started finding my own photos or building rooms in stuff like the sims or minecraft but something and then trying to apply that same lighting.

but what i'm getting at in all this long drawn out response is it's better to look at like actual people art or photos and so on to study from coz the ai is just teaching mistakes so while i can get the excuse of it being more convenient it's still a bad teacher coz it's also giving a bunch of errors which a person then copies and wont know till its pointed out if ever coz the focus will probably be more on the use of ai than the attempt at learning and failing

[ShidoDraws]

But it can also be a slippery slope because that's not far off from condemning fanfiction, fanart, drawing inspiration from existing works, etc. Artists don't consent to their work being used for AI, but they also don't consent to fanart being made in their style (unless they expressly support it).

About that, making fan contents and inspired works is not illegal. Now, if we're talking about monetizing fan content, that's whole other legal issue/debate. As for AI datasets being a huge problem because people's work was stolen. One could argue that when people sign up for the website, they've given permission for those sites to use/distribute their work. However, the first internet wide scrapes were secretly ripped from websites "bypassing" the Terms and Agreements users has initially signed up for. Microsoft and Alphabet is also a whole other issue; I do not know the legal stuff but for sure should raise privacy concern. The ones whose work that aren't on the internet is being scrapped from their cloud and those settings being opt in for Windows 11 users. With the progression of these issues, I don't even know if we'll be able to actually own a computer in the near future because Nvidia will push these companies to turn computers, phones, maybe even cars into subscription-based products so they can afford to keep the AI data centers running. I do not think that the expensive part is building these data centers, it will be the maintenance part that will be expensive. If you look at the map that shows where the data centers are, they are sparsely placed in the US. It's a strategic placing, so when things around it becomes expensive and people lose their homes, it wouldn't cause a whole lot of noise. And I call homelessness in a quiet chaos.

Anyway, fan contents and inspirational contents have found their place in this world because it doesn't steal people's jobs. If anything, they make jobs, self-employment is huge because of it. Generative AI on the other hand, are replacing people because employers believe it's cheaper and are more obedient. However, for the sake of the argument, let's pretend generative AI isn't stealing people's jobs and everyone is using it to keep an income, employed or self-employed. In order for people to actually properly compensate these tech companies for the technology, we'd basically be paying a monthly of 2,000 USD per person for it to be worth it for them, and I'm being generous with the pricing. Truthfully, most people don't have that kind of money. Tech companies don't want people to know about this, that is one of the reasons why people call it a bubble. Once the price to use the technology goes up, some people would have no choice but to keep paying for it because they've bult their livelihood with/around generative AI. So, in the end you have to learn how to draw/write or commission humans if you want something cheap.

[JWZagst]

Fair enough. I really don't have anything more to add to the conversation, so I will see myself out XD.

[stiles]