7 ms·

Why do people keep bringing up copyright anyway? It seems pretty clear that the images being generated by StableDiffusion are transformative so it's protected u

Why do people keep bringing up copyright anyway? It seems pretty clear that the images being generated by StableDiffusion are transformative so it's protected under Fair Use. Am I missing something here?

imgabe 4y ago
Fair use is a defense. You can still be sued for fair use, you have to go to court and prove to a judge or jury that your work constitutes fair use.
- LewisVerstappen 4y ago
  Yes but we have a common law system and there's already tons of precedent that training AI systems is transformative. It's also quite obvious just by looking at the generated images that it's clearly transformative. The images generated are unique and you can't trace the original copyrighted image from what's generated. You really don't need a judge to see that Fair Use covers Stable Diffusion.
  - Natsu 4y ago
    You can beat the rap but you can't beat the ride. Federal lawsuits are not cheap and the default is that you pay your own costs, you have to win the argument that you should get court costs & attorney's fees.
  - rhdunn 4y ago
    What happens if you give an image prompt like "mona lisa", "daffodils van gogh", or similar designed to describe an image the model was trained on. Will it generate that image? Or for written works, start with a sentance from a copyrighted work, or part of licensed code. Will it start reproducing that work word for word (like code pilot can do with the GPL license)? Getting these to generate copies of GPL'd, company owned, or other code with restrictions can lead to complex issues for the person/company using that code. Or likewise if a story contains significant elements of copyrighted works; worse if the works have trademarked elements.
    - slyall 4y ago
      What do you picture when you see words like "Elvis Presley" , "Game of Thrones" or "The Google logo"? Explain how you are not a copyright violating machine.
      - sarchertech 4y ago
        I absolutely would be if I took copyrighted images out of my head and distributed them.
  - criddell 4y ago
    > The images generated are unique and you can't trace the original copyrighted image from what's generated. That might not always be true. I've gotten some results back that had the Getty watermark on them and others with the artist's signature. Unless the AI is adding that to images that never had one before (which might be a trademark issue), then you might be able to determine the provenance of the image components. Is there something equivalent to the yellow dots printers add to their output that would survive the AI transformation?
- PeterisP 4y ago
  Well, you can still be sued even if the reason is almost totally bogus, and you have to go to court and make a stand that the reason is bogus. Doing only 100% legal things won't protect you from being sued.
throwaway1851 4y ago
There are really significant, novel copyright issues implicated by these large generative models trained on other people’s IP. If you take a step back, you can see that there are different ways to frame what is happening. One frame is: “Defendant built an algorithm that memorized features of Plaintiff’s IP. Defendant’s algorithm recombines parts of those features in order to produce works in the same domain that compete with Plaintiff’s work, all without Plaintiff’s consent.” Bear in mind that copyright holders are among the most litigious out there. If generative art becomes as big a deal as some people expect, they will have every incentive to use their huge litigation budgets to claim a piece of the action. Because the law develops very slowly, the legal process has not yet had the occasion to really evaluate what transformative use means in this novel context. I’m personally interested in seeing where things go, but it’s going to be a while before we know where the law is headed.
- klipt 4y ago
  > “Defendant built an algorithm that memorized features of Plaintiff’s IP. Defendant’s algorithm recombines parts of those features in order to produce works in the same domain that compete with Plaintiff’s work, all without Plaintiff’s consent.” The fun part is, this is how human artists learn too.
  - throwaway1851 4y ago
    They absolutely do. What they don’t do is mechanistically clone compressed mathematical representations of input data. The human part of the creative process could very well be a distinguishing feature, legally.
    - nilsbunger 4y ago
      It will be hard to tell whether an image was created by a “human creative process” or AI. Will the same image be legal if a human made it but not if it was created by Stable Diffusion? How will someone even know, short of a legal discovery process?
Toadtoad 4y ago
Whether a work is transformative or not isn't the only factor that goes towards a fair use ruling. Bring transformative is just a part of the "purpose and character of use", which is a single factor weighed alongside the nature of the original work, the amount and substance used (if the "heart of the work" was copied), and the effect the derivative has on the market for the original work. It's much more complicated here, and will have to be ruled on a case-by-case basis, as far as I know, and copyright holders could potentially make many cases using the other factors that a use isn't fair. I do think it should be considered fair use, but as the law currently stands, it does appear to be more complex, at least from my point of view. Note: I am absolutely not a legal expert.
jtolmar 4y ago
The new technology looks like it could threaten artists' livelihoods by replicating something about their work. The old legal framework for protecting artists' livelihoods from photocopiers etc was copyright. The goal of making sure artists can both make art and eat is still relevant, so people reach for the tool that used to make sure that was possible.
- gernb 4y ago
  If it really does make it hard for artists to eat from selling art, that to me is just sour grapes. I mean there use to be 1000s or 10s of 1000s of liveries to lock up your horse. There's no goal to keep horse keepers so they can still keep horses and eat. Instead, their industry mostly disappeared. I'm not heartless but at the point their services are no longer needed that's just the way it is. Cloth weavers replaced by the loom. I don't think SD will replace most top artists for now. It's hard for me to believe SD is going to come up with images like those from top concept artists. But I can imagine SD replacing lots of situations, like maybe stock photography, when you can just ask the AI to draw "people in front of whiteboard discussing sales chart"
- cma 4y ago
  We need a different tool than artificial scarcity. Both for copyrights and patents. Government funded prize systems and patronage systems with a lot of mechanisms for citizens to choose what to fund seems ideal.
astrange 4y ago
Online artists esp. fanartists have a strict moral system with rules like “credit the original artist” that isn’t based on actual laws, so they’re upset about this.
dkjaudyeqooe 4y ago
I think that's wishful thinking. It's objectively a mechanical derivation based on many copyrighted works. In many cases it's going to reproduce some works with not that much transformation. What's a similar thing that has come before this? I can't think of any, this is very novel. You'd want to wait for some rulings before you jump to conclusions.
- elevaet 4y ago
  It might be similar to some forms of (human composed) sample-heavy music which use bits and pieces from many different songs to create something entirely original. As far as I understand it, this is still considered copyright infringement in most IP law systems. (If the samples aren't cleared)
grumbel 4y ago
Fair Use is a defense, not a right. Just being transformative isn't enough here, that's just one of many different factors that needs to be checked. Furthermore Fair Use includes evaluating the effect an infringement has upon original work's value. So when your image generator keeps spitting out Gettyimages watermarks, while you are building a service that is in direct competition with Gettyimages for stock images, there is an argument to be made that Fair Use really doesn't apply here. As what you are doing is essentially stealing Gettyimages' work, AI laundering it and selling it back to their previous customers. With StableDiffusion a Fair Use defense might have an easier time, as the results are released to the public. But it's still not exactly clear cut. If you type in "Mona Lisa", you'll still get something that looks like a copy of the Mona Lisa, not like an original work.

imgabe 4y ago

Fair use is a defense. You can still be sued for fair use, you have to go to court and prove to a judge or jury that your work constitutes fair use.

LewisVerstappen 4y ago
Yes but we have a common law system and there's already tons of precedent that training AI systems is transformative. It's also quite obvious just by looking at the generated images that it's clearly transformative. The images generated are unique and you can't trace the original copyrighted image from what's generated. You really don't need a judge to see that Fair Use covers Stable Diffusion.
- Natsu 4y ago
  You can beat the rap but you can't beat the ride. Federal lawsuits are not cheap and the default is that you pay your own costs, you have to win the argument that you should get court costs & attorney's fees.
- rhdunn 4y ago
  What happens if you give an image prompt like "mona lisa", "daffodils van gogh", or similar designed to describe an image the model was trained on. Will it generate that image? Or for written works, start with a sentance from a copyrighted work, or part of licensed code. Will it start reproducing that work word for word (like code pilot can do with the GPL license)? Getting these to generate copies of GPL'd, company owned, or other code with restrictions can lead to complex issues for the person/company using that code. Or likewise if a story contains significant elements of copyrighted works; worse if the works have trademarked elements.
  - slyall 4y ago
    What do you picture when you see words like "Elvis Presley" , "Game of Thrones" or "The Google logo"? Explain how you are not a copyright violating machine.
    - sarchertech 4y ago
      I absolutely would be if I took copyrighted images out of my head and distributed them.
- criddell 4y ago
  > The images generated are unique and you can't trace the original copyrighted image from what's generated. That might not always be true. I've gotten some results back that had the Getty watermark on them and others with the artist's signature. Unless the AI is adding that to images that never had one before (which might be a trademark issue), then you might be able to determine the provenance of the image components. Is there something equivalent to the yellow dots printers add to their output that would survive the AI transformation?
PeterisP 4y ago
Well, you can still be sued even if the reason is almost totally bogus, and you have to go to court and make a stand that the reason is bogus. Doing only 100% legal things won't protect you from being sued.

throwaway1851 4y ago

There are really significant, novel copyright issues implicated by these large generative models trained on other people’s IP. If you take a step back, you can see that there are different ways to frame what is happening. One frame is: “Defendant built an algorithm that memorized features of Plaintiff’s IP. Defendant’s algorithm recombines parts of those features in order to produce works in the same domain that compete with Plaintiff’s work, all without Plaintiff’s consent.” Bear in mind that copyright holders are among the most litigious out there. If generative art becomes as big a deal as some people expect, they will have every incentive to use their huge litigation budgets to claim a piece of the action. Because the law develops very slowly, the legal process has not yet had the occasion to really evaluate what transformative use means in this novel context. I’m personally interested in seeing where things go, but it’s going to be a while before we know where the law is headed.

klipt 4y ago
> “Defendant built an algorithm that memorized features of Plaintiff’s IP. Defendant’s algorithm recombines parts of those features in order to produce works in the same domain that compete with Plaintiff’s work, all without Plaintiff’s consent.” The fun part is, this is how human artists learn too.
- throwaway1851 4y ago
  They absolutely do. What they don’t do is mechanistically clone compressed mathematical representations of input data. The human part of the creative process could very well be a distinguishing feature, legally.
  - nilsbunger 4y ago
    It will be hard to tell whether an image was created by a “human creative process” or AI. Will the same image be legal if a human made it but not if it was created by Stable Diffusion? How will someone even know, short of a legal discovery process?

Toadtoad 4y ago

Whether a work is transformative or not isn't the only factor that goes towards a fair use ruling. Bring transformative is just a part of the "purpose and character of use", which is a single factor weighed alongside the nature of the original work, the amount and substance used (if the "heart of the work" was copied), and the effect the derivative has on the market for the original work. It's much more complicated here, and will have to be ruled on a case-by-case basis, as far as I know, and copyright holders could potentially make many cases using the other factors that a use isn't fair. I do think it should be considered fair use, but as the law currently stands, it does appear to be more complex, at least from my point of view. Note: I am absolutely not a legal expert.

jtolmar 4y ago

The new technology looks like it could threaten artists' livelihoods by replicating something about their work. The old legal framework for protecting artists' livelihoods from photocopiers etc was copyright. The goal of making sure artists can both make art and eat is still relevant, so people reach for the tool that used to make sure that was possible.

gernb 4y ago
If it really does make it hard for artists to eat from selling art, that to me is just sour grapes. I mean there use to be 1000s or 10s of 1000s of liveries to lock up your horse. There's no goal to keep horse keepers so they can still keep horses and eat. Instead, their industry mostly disappeared. I'm not heartless but at the point their services are no longer needed that's just the way it is. Cloth weavers replaced by the loom. I don't think SD will replace most top artists for now. It's hard for me to believe SD is going to come up with images like those from top concept artists. But I can imagine SD replacing lots of situations, like maybe stock photography, when you can just ask the AI to draw "people in front of whiteboard discussing sales chart"
cma 4y ago
We need a different tool than artificial scarcity. Both for copyrights and patents. Government funded prize systems and patronage systems with a lot of mechanisms for citizens to choose what to fund seems ideal.

astrange 4y ago

Online artists esp. fanartists have a strict moral system with rules like “credit the original artist” that isn’t based on actual laws, so they’re upset about this.

dkjaudyeqooe 4y ago

I think that's wishful thinking. It's objectively a mechanical derivation based on many copyrighted works. In many cases it's going to reproduce some works with not that much transformation. What's a similar thing that has come before this? I can't think of any, this is very novel. You'd want to wait for some rulings before you jump to conclusions.

elevaet 4y ago
It might be similar to some forms of (human composed) sample-heavy music which use bits and pieces from many different songs to create something entirely original. As far as I understand it, this is still considered copyright infringement in most IP law systems. (If the samples aren't cleared)

grumbel 4y ago

Fair Use is a defense, not a right. Just being transformative isn't enough here, that's just one of many different factors that needs to be checked. Furthermore Fair Use includes evaluating the effect an infringement has upon original work's value. So when your image generator keeps spitting out Gettyimages watermarks, while you are building a service that is in direct competition with Gettyimages for stock images, there is an argument to be made that Fair Use really doesn't apply here. As what you are doing is essentially stealing Gettyimages' work, AI laundering it and selling it back to their previous customers. With StableDiffusion a Fair Use defense might have an easier time, as the results are released to the public. But it's still not exactly clear cut. If you type in "Mona Lisa", you'll still get something that looks like a copy of the Mona Lisa, not like an original work.