5 ms·

If a human learns to draw by studying the works of other artists, is that infringement?

by Razengan 3y ago

xyzzy123 3y ago
If an algorithm processes a bunch of copyrighted data into a different form, is that not a derivative work?
- judge2020 3y ago
  only if the output is definitively a derivative work of an existing work. If it's the culmination of a thousand photos, it would be akin to a human spending a month studying a certain historical artists' art and painting a modern Tesla in their style.
jackvalentine 3y ago
It could be, if the human uses their skills to draw infringing outputs!
- a_bonobo 3y ago
  hence the entire Andy Warhol case! >Echoing a recommendation from Joe Biden’s administration, the supreme court focused on the specific use that allegedly infringed Goldsmith’s copyright – a license of Warhol’s work to Condé Nast – and said it was not transformative because it served the same commercial purpose as Goldsmith’s photo: to depict Prince in a magazine. https://www.theguardian.com/artanddesign/2023/may/18/andy-warhol-copyright-prince-paintings-lawsuit https://www.theguardian.com/artanddesign/2023/may/18/andy-wa...
- brookst 3y ago
  So why are we even arguing about this new form of copyright that restricts who can learn from a work? If an AI model produces an infringing work and someone distributes it, sue them like you would a human doing the same thing. This idea of training being a violation of copyright is insane.
  - mattigames 3y ago
    No, is not insane at all, a computer can generate millions of derivate copies in a second, a human has limited time for learning and for making artwork, so it doesn't make sense to compare them, not in the slightest. If suddenly there was a race of humans who could read and retain entire books in a matter of seconds (and other similar feats) the implications would be almost the same, as it would break basic expectations of time and personal resources that our society relies on, and a lot of people would be rightfully worried about such people entering the workforce.
    - Solvency 3y ago
      Computers can do all sorts of things humans can't at inhuman scale and speed. Why is AI leveraging the same training data humans do inherently bad just because it can do it better and faster than humans?
      - mattigames 3y ago
        Yeah, I'm a programmer, I'm aware of all that by definition. The issue here it's getting to he point where is inevitable that it will replace a huge amount of job positions in a tiny frame of time in ways society is not prepared for, if you think the amount of jobs at risk is anywhere in a similar scale we have seen before you haven't been paying attention; of course the flaw here lies on society's itself and what it means to earn a living, and not really over ML researchers or Adobe or any other company creating those AIs.
    - zirgs 3y ago
      Adobe trained their own model only on public domain images and on images that they own. Do you think that they should not be allowed to do that?
      - mattigames 3y ago
        That's a different discussion and should have wide ramifications with way more nuance that "allowed" or "not allowed", if suddenly tomorrow Adobe or any other were able to create a model that is able to replace any existing job position how we should it be managed? The societal ramifications are complex, is not binary and should be discussed by politicians and intellectuals, not by corp marketing teams.
krapp 3y ago
No. But we're not talking about humans, we're talking about software intended to put the artists whose work it was trained on - without permission or compensation - out of business, or which is at least capable of devaluing their work with its ability to copy styles (literally by name.) Even if one could argue (as I probably have) that AI is more than simply a "stochastic parrot," a philosophical argument about whether it resembles a primitive mind (and therefore whether its output can be considered "art" in any form) doesn't answer a legal argument about copyright infringement. Humans are allowed to be inspired by other humans, software used by other humans is not. Also you shouldn't delete your comment just because it gets downvotes and repost it. Bad form.
- ethbr0 3y ago
  The bigger argument is in the form of cost of reproduction. A human has finite time and output volume. A machine has no such limits, only compute/memory/network/disk. Ergo, gray areas where we could afford a human the benefit of doubt ("What? They can paint 4 in-the-style-of-Rembrandts?") produce substantially different results with machines ("Here are a million in-the-style-of-Rembrandts, produced in 30 minutes"). Whether we should permit or limit machine art is a discussion worth having -- but human copying vs machine copying is clearly a very different discussion.
  - pixl97 3y ago
    Right, this is why we should burn the combines and go back to harvesting wheat by scythe, or we should dig ditches with spoons instead of using an excavator... The problem here is we want the system we currently have to keep working when there is really no possible way that can happen in a world of AGI. IP is already a horribly broken system that can lead to absurdities. While making IP stronger will protect people like artists (maybe) it will provide far more protection to monied corporations that can afford to buy up works, and then feed them into their AI. In the pathological cases you end up with stories like 'The Right To Read'. In a world where particular things can be created in almost infinite amounts, why will we go to such efforts to ensure they are only created in limited supply?
torginus 3y ago
So if you're making the argument that AI is a human when it comes to the copyright, does that mean that its output is copyrightable just like a humans? Does that mean that OpenAI/SD whoever now owns the work created in such a way? It would logically follow from your reasoning.
maeln 3y ago
For what should be pretty obvious reason, most society and laws will treat humans and corporation and their output differently.
amelius 3y ago
Why are you assuming that deep learning and (human) learning are the same thing?
- melagonster 3y ago
  because he trust Human just a larger matrix.

If a human learns to draw by studying the works of other artists, is that infringement?

xyzzy123 3y ago

If an algorithm processes a bunch of copyrighted data into a different form, is that not a derivative work?

judge2020 3y ago
only if the output is definitively a derivative work of an existing work. If it's the culmination of a thousand photos, it would be akin to a human spending a month studying a certain historical artists' art and painting a modern Tesla in their style.

jackvalentine 3y ago

It could be, if the human uses their skills to draw infringing outputs!

a_bonobo 3y ago
hence the entire Andy Warhol case! >Echoing a recommendation from Joe Biden’s administration, the supreme court focused on the specific use that allegedly infringed Goldsmith’s copyright – a license of Warhol’s work to Condé Nast – and said it was not transformative because it served the same commercial purpose as Goldsmith’s photo: to depict Prince in a magazine. https://www.theguardian.com/artanddesign/2023/may/18/andy-warhol-copyright-prince-paintings-lawsuit https://www.theguardian.com/artanddesign/2023/may/18/andy-wa...
brookst 3y ago
So why are we even arguing about this new form of copyright that restricts who can learn from a work? If an AI model produces an infringing work and someone distributes it, sue them like you would a human doing the same thing. This idea of training being a violation of copyright is insane.
- mattigames 3y ago
  No, is not insane at all, a computer can generate millions of derivate copies in a second, a human has limited time for learning and for making artwork, so it doesn't make sense to compare them, not in the slightest. If suddenly there was a race of humans who could read and retain entire books in a matter of seconds (and other similar feats) the implications would be almost the same, as it would break basic expectations of time and personal resources that our society relies on, and a lot of people would be rightfully worried about such people entering the workforce.
  - Solvency 3y ago
    Computers can do all sorts of things humans can't at inhuman scale and speed. Why is AI leveraging the same training data humans do inherently bad just because it can do it better and faster than humans?
    - mattigames 3y ago
      Yeah, I'm a programmer, I'm aware of all that by definition. The issue here it's getting to he point where is inevitable that it will replace a huge amount of job positions in a tiny frame of time in ways society is not prepared for, if you think the amount of jobs at risk is anywhere in a similar scale we have seen before you haven't been paying attention; of course the flaw here lies on society's itself and what it means to earn a living, and not really over ML researchers or Adobe or any other company creating those AIs.
  - zirgs 3y ago
    Adobe trained their own model only on public domain images and on images that they own. Do you think that they should not be allowed to do that?
    - mattigames 3y ago
      That's a different discussion and should have wide ramifications with way more nuance that "allowed" or "not allowed", if suddenly tomorrow Adobe or any other were able to create a model that is able to replace any existing job position how we should it be managed? The societal ramifications are complex, is not binary and should be discussed by politicians and intellectuals, not by corp marketing teams.

krapp 3y ago

No. But we're not talking about humans, we're talking about software intended to put the artists whose work it was trained on - without permission or compensation - out of business, or which is at least capable of devaluing their work with its ability to copy styles (literally by name.) Even if one could argue (as I probably have) that AI is more than simply a "stochastic parrot," a philosophical argument about whether it resembles a primitive mind (and therefore whether its output can be considered "art" in any form) doesn't answer a legal argument about copyright infringement. Humans are allowed to be inspired by other humans, software used by other humans is not. Also you shouldn't delete your comment just because it gets downvotes and repost it. Bad form.

ethbr0 3y ago
The bigger argument is in the form of cost of reproduction. A human has finite time and output volume. A machine has no such limits, only compute/memory/network/disk. Ergo, gray areas where we could afford a human the benefit of doubt ("What? They can paint 4 in-the-style-of-Rembrandts?") produce substantially different results with machines ("Here are a million in-the-style-of-Rembrandts, produced in 30 minutes"). Whether we should permit or limit machine art is a discussion worth having -- but human copying vs machine copying is clearly a very different discussion.
- pixl97 3y ago
  Right, this is why we should burn the combines and go back to harvesting wheat by scythe, or we should dig ditches with spoons instead of using an excavator... The problem here is we want the system we currently have to keep working when there is really no possible way that can happen in a world of AGI. IP is already a horribly broken system that can lead to absurdities. While making IP stronger will protect people like artists (maybe) it will provide far more protection to monied corporations that can afford to buy up works, and then feed them into their AI. In the pathological cases you end up with stories like 'The Right To Read'. In a world where particular things can be created in almost infinite amounts, why will we go to such efforts to ensure they are only created in limited supply?

torginus 3y ago

So if you're making the argument that AI is a human when it comes to the copyright, does that mean that its output is copyrightable just like a humans? Does that mean that OpenAI/SD whoever now owns the work created in such a way? It would logically follow from your reasoning.

maeln 3y ago

For what should be pretty obvious reason, most society and laws will treat humans and corporation and their output differently.

amelius 3y ago

Why are you assuming that deep learning and (human) learning are the same thing?

melagonster 3y ago
because he trust Human just a larger matrix.