...This is the one argument about gen AI that I've never understood.
A) Why would I want a reference from something that doesn't even know what reality is? Wouldn't the best references come from the real world, not a machine that just guesses what it probably looks like?
It would be like choosing someone to interview about the fashion industry, and instead of looking for a buyer or a designer or even just a design student, you pick some rando who just has a lot of cute outfits on their Pinterest board. Like, maybe they do know something valuable...but out of all the options available, they're the least likely to, so why are they first choice??
B) If [insert pose] isn't available online, where is the AI getting its reference info from? ^^; Am I asking it to just wholesale hallucinate something for me now? Why would I want it to do that; wouldn't that just increase the likelihood of it giving me something that doesn't make logical sense...?
C) Just look up 'artist reference libraries' and pick a website. There are SO many. REAL people have been working hard to help their fellow artists with this problem long before gen AI became a thing, and they actually try come up with refs that are interesting, not just the lowest-common-denominator slop that AI tends to spit out.
Duchess Celestia has a whole video (from last year, so this is with knowledge of the AI-pocalypse) that lists a ton of them: https://www.youtube.com/watch?v=pg0oDgXQB6U
There is no need to rely on AI for this, unless your actual problem is that you're just too lazy to find what you want from a quality resource. In which case I would argue; why bother drawing at all then? Just let the AI do the art for you; then you won't even need a reference. Kill two birds with one stone. ¯_(ツ)_/¯
Another thing people keep saying without sources...if these models exist, please list them. I think people need to know about them. Give company names; give dataset names so we can verify.
I don't doubt that ethical gen AI datasets/use cases exist, but you can't use that as a counter-argument without proof. Also, counter-counter-argument: people's disdain is still valid if the biggest datasets that see the most use are blatantly unethical. And so far that's still the case.
Yes and no.
Obviously human-made work will gain value if it's less incentivized/available; that's how scarcity works. But it won't be much if people stop caring about it, which is where the push is being made.
I think the main difference between gen AI and other "innovations" throughout history is that the others felt no need to lie about what they were or hide their usage. Photographers were proud to use cameras and sell people on the differences between their art and traditional portrait painting. Early automobile enthusiasts (to give another example) were excited about their craft, no matter how loud and messy and dangerous it was at first. They approached their hobby with the same energy people gave to horses.
Generative AI, on the other hand, is not an art or a craft or a hobby, and 9 times outta 10 we have to "find out" that it's being used, how, and where. Even when its loudest supporters admit that it's being used, they're always dodging questions about the specifics, claiming that no one will lose their jobs (spoiler: they will), and that everything about the quality of the artistic pipeline and end product will remain the same...which begs the question, what's the point of even using it, then...?
The real answer is that it saves the most annoying, soulless people on earth time and money, and lets them roleplay the concept of imagination, and that's pretty much it. =T But they don't want to admit that; they want to sell everyone on the idea that it will make our lives better and more convenient while simultaneously not changing anything that matters. The whole entire point is just for us to look around at the world as it exists right now, mentally replace every semblance of human input with LLM output, and not notice or care about the difference.
So with that in mind...if these people get the future they want, why should anyone in it care about human-made art? It's supposed to be indistinguishable from AI art anyway; all they'll learn from finding out that you made it is that it took 10x longer than it should have. Even if they like watching the process of creating it, they can always watch an AI pretend to create the same thing...your involvement will only matter to someone who is already an artist. To everyone else, it'll be little more than a neat piece of trivia.
Of course, I don't think ^that will happen; realistically I think AI-generated images will be Clip Art on steroids at best-- a marker of low quality and the bare minimum of effort; appropriate for some cases but heavily looked down upon in others. And Clip Art/stockphotos didn't magically make artists more valuable, so I doubt this will either.
On the contrary, it will probably just eat up more entry-level jobs and make the job market smaller, basically eliminating the gains we made when digital media exploded in the 90's and 00's, and sending us back to the 20th century in terms of career viability. A change that most normal people weren't aware of anyway. ^^; So.
