4 ms·

I don't think the question is really about whether AI art is real art. (But it could be about that, as I'm not familiar with commercial cons in the US.) Some y

I don't think the question is really about whether AI art is real art. (But it could be about that, as I'm not familiar with commercial cons in the US.) Some years ago, around the time I became aware that AI art is a thing, the artist scene around Finnish cons had already decided to ban it. And the reason was obvious, as the same people are also very eager to police others who might be selling pirated products. They don't care legal constructs such as intellectual property. They don't really care about economic constructs such as copyright. What they care about are authors' moral rights. If the model was trained without obtaining a permission from the authors of every work in the training data, they think using the model to create art is immoral.

t0bia_s 9mo ago
AI generative art doesn't exist by definition. You cannot generate art. Actually we have a term for this - kitsch.
- UncleMeat 9mo ago
  I think that you can probably do some interesting things. Sol Lewitt made art that was just instructions to be interpreted by another human. But the medium needs to be the instructions rather than the isolated output of the machine.
rickydroll 9mo ago
> What they care about are authors' moral rights. If the model was trained without obtaining permission from the authors of every work in the training data, they think using the model to create art is immoral. Art is not created in isolation. It is a result of the artist's exposure (aka training), both intentional and incidental. If an artist wants an AI model to get permission before training on their work, then the artist should get permission from all the artists they were exposed to that shaped their artistic expression. It's training and copying all the way down.
- spwa4 9mo ago
  The problem is that if this argument is allowed to stand, art, as a human endeavor will shrink 99% or maybe even 100%. Oh and this happens in a very underhanded way. Courts, governments and companies (including OpenAI and others) demand copyright is respected by humans. They impose great penalties when humans cheat, and then this happens: https://torrentfreak.com/nvidia-contacted-annas-archive-to-secure-access-to-millions-of-pirated-books/ https://torrentfreak.com/nvidia-contacted-annas-archive-to-s... https://torrentfreak.com/authors-accuse-openai-of-using-pirate-sites-to-train-chatgpt-230630/ https://torrentfreak.com/authors-accuse-openai-of-using-pira... https://torrentfreak.com/meta-torrented-over-81-tb-of-data-through-annas-archive-despite-few-seeders-250206/ https://torrentfreak.com/meta-torrented-over-81-tb-of-data-t... If these companies were forced to abide by the rules courts impose on humans, they would have to buy billions worth of books. But of course, "that's not how copyright works". Of course, these companies ARE using copyright to avoid reciprocating: https://openai.com/policies/row-terms-of-use/ https://openai.com/policies/row-terms-of-use/ So this is yet another "rules for thee, not for me" situation involving companies worth billions of dollars. A situation that's really hurting people's livelihoods ...
  - rickydroll 9mo ago
    I can't disagree with your AI doomerism perspective. I firmly believe that AI companies should buy one copy of whatever work they use for training. While this won't provide the never-ending royalty stream on copyrighted material that corporations strive for, it would foster the mindset that AI companies must pay society in some way. And I truly think that if AI companies are going to train on all the knowledge in the world, their profits should go back to everyone in the world. i.e., LLM models are a public good. I have an almost unshakable conviction that LLM-type AI systems should become a repository of all human knowledge. When LLMs give you an answer, you should be able to ask, what are the sources behind your answer? People won't do this, but curious, wanting-to-learn people will. Which leads to one of the important questions. How do you keep people curious?
    - spwa4 9mo ago
      But all these companies violated that on a massive scale. It's done. They're not paying. Oh, and when asked what the consequences are for people doing illegal downloading, ChatGPT helpfully answers: > About $750 to $30,000 per copyrighted work > Can go up to $150,000 per work if it’s considered willful ... it was definitely willful. And these are amounts that would bankrupt even OpenAI. But I guess only you and me will have to pay these sorts of amounts, not big companies ...
jezzamon 9mo ago
To clarify, I'm not saying AI created stuff can't be art, I'm saying that someone that enters a text prompt is not the creator of the AI's output
- thedevilslawyer 9mo ago
  Technically, it's a legal grey area, and currently any image by AI can be considered public domain. This is a good change in society towards protectionist IP, which was long due for fixing, but was never done.
Suppafly 9mo ago
>If the model was trained without obtaining a permission from the authors of every work in the training data, they think using the model to create art is immoral. Honestly, they aren't even that logical. Most of the anti-ai rhetoric is anti any usage, even that which is fully licensed.

t0bia_s 9mo ago

AI generative art doesn't exist by definition. You cannot generate art. Actually we have a term for this - kitsch.

UncleMeat 9mo ago
I think that you can probably do some interesting things. Sol Lewitt made art that was just instructions to be interpreted by another human. But the medium needs to be the instructions rather than the isolated output of the machine.

rickydroll 9mo ago

> What they care about are authors' moral rights. If the model was trained without obtaining permission from the authors of every work in the training data, they think using the model to create art is immoral. Art is not created in isolation. It is a result of the artist's exposure (aka training), both intentional and incidental. If an artist wants an AI model to get permission before training on their work, then the artist should get permission from all the artists they were exposed to that shaped their artistic expression. It's training and copying all the way down.

spwa4 9mo ago
The problem is that if this argument is allowed to stand, art, as a human endeavor will shrink 99% or maybe even 100%. Oh and this happens in a very underhanded way. Courts, governments and companies (including OpenAI and others) demand copyright is respected by humans. They impose great penalties when humans cheat, and then this happens: https://torrentfreak.com/nvidia-contacted-annas-archive-to-secure-access-to-millions-of-pirated-books/ https://torrentfreak.com/nvidia-contacted-annas-archive-to-s... https://torrentfreak.com/authors-accuse-openai-of-using-pirate-sites-to-train-chatgpt-230630/ https://torrentfreak.com/authors-accuse-openai-of-using-pira... https://torrentfreak.com/meta-torrented-over-81-tb-of-data-through-annas-archive-despite-few-seeders-250206/ https://torrentfreak.com/meta-torrented-over-81-tb-of-data-t... If these companies were forced to abide by the rules courts impose on humans, they would have to buy billions worth of books. But of course, "that's not how copyright works". Of course, these companies ARE using copyright to avoid reciprocating: https://openai.com/policies/row-terms-of-use/ https://openai.com/policies/row-terms-of-use/ So this is yet another "rules for thee, not for me" situation involving companies worth billions of dollars. A situation that's really hurting people's livelihoods ...
- rickydroll 9mo ago
  I can't disagree with your AI doomerism perspective. I firmly believe that AI companies should buy one copy of whatever work they use for training. While this won't provide the never-ending royalty stream on copyrighted material that corporations strive for, it would foster the mindset that AI companies must pay society in some way. And I truly think that if AI companies are going to train on all the knowledge in the world, their profits should go back to everyone in the world. i.e., LLM models are a public good. I have an almost unshakable conviction that LLM-type AI systems should become a repository of all human knowledge. When LLMs give you an answer, you should be able to ask, what are the sources behind your answer? People won't do this, but curious, wanting-to-learn people will. Which leads to one of the important questions. How do you keep people curious?
  - spwa4 9mo ago
    But all these companies violated that on a massive scale. It's done. They're not paying. Oh, and when asked what the consequences are for people doing illegal downloading, ChatGPT helpfully answers: > About $750 to $30,000 per copyrighted work > Can go up to $150,000 per work if it’s considered willful ... it was definitely willful. And these are amounts that would bankrupt even OpenAI. But I guess only you and me will have to pay these sorts of amounts, not big companies ...

jezzamon 9mo ago

To clarify, I'm not saying AI created stuff can't be art, I'm saying that someone that enters a text prompt is not the creator of the AI's output

thedevilslawyer 9mo ago
Technically, it's a legal grey area, and currently any image by AI can be considered public domain. This is a good change in society towards protectionist IP, which was long due for fixing, but was never done.

Suppafly 9mo ago

>If the model was trained without obtaining a permission from the authors of every work in the training data, they think using the model to create art is immoral. Honestly, they aren't even that logical. Most of the anti-ai rhetoric is anti any usage, even that which is fully licensed.