10 ms·

> But the latest and greatest software trend–generative AI–is in danger of being swallowed up by copyright law. About time. > If the AI industry is to survive

> But the latest and greatest software trend–generative AI–is in danger of being swallowed up by copyright law. About time. > If the AI industry is to survive, we need a clear legal rule that neural networks, and the outputs they produce, are not presumed to be copies of the data used to train them. But they are compressed lossy copies of all that data! That's the whole point of noise/denoise functions that neural networks are based upon. The whole mathematical foundation of training a neural network is "teaching" it how to recognize and/or create copies of data stored in the training set. > Otherwise, the entire industry will be plagued with lawsuits that will stifle innovation and only enrich plaintiff’s lawyers. When you're willingly breaking already established law en masse for profit in hope no one cares enough, be it copyright law or any other, you're not an "innovator", you're a criminal. The fact that you're a tech giant or a Bay startup doesn't matter in this regard; the only thing that matters is the notable amount of time required for the justice system to catch up with your novel tools for laundering intellectual property.

tomrod 4y ago
Transgressors, perhaps, but not criminals. AI, especially generative AI, are not lossy copies but rather synthesizers. If anything, your comment highlights the need to evolve IP norms and laws.
thewakalix 4y ago
How are you using the word “copy”? It doesn’t seem to match the standard meaning. For instance, most people would not consider a brief summary of a movie’s plot to be a “copy” of that movie, or protected under copyright.
- phoe-krk 4y ago
  If you have an image, then train a neural network on that image, then use the neural network to reconstruct that image in detail, then the NN by definition contains enough information used to reconstruct that image - hence, a copy. With NNs trained on thousands or millions of data entries, this concept becomes fuzzy in the same way as you described - a short summary likely wouldn't be considered a copy, just like a 64x64 generated thumbnail wouldn't be considered in the same way a 4096x4096 hi-res image.
  - Krasnol 4y ago
    > If you have an image, then train a neural network on that image, then use the neural network to reconstruct that image in detail I haven't seen that happening since the discussion started. Most of the complains I saw aimed at things like "it stole my style" not "it reproduced my art". Do you have any examples?
    - greysphere 4y ago
      It's more, 'this product is profiting from my labor without my consent (ie paying me).' In music you aren't allowed to use the same notes, even if you played them on a trumpet with a swing beat, while the source was on the piano very staccato. While we don't have the same vocabulary for art, it's not unreasonable to expect similar protections.
      - wvenable 4y ago
        What you describe for music is already outrageous -- why do you think that needs to be extended to everything? https://www.vice.com/en/article/wxepzw/musicians-algorithmically-generate-every-possible-melody-release-them-to-public-domain https://www.vice.com/en/article/wxepzw/musicians-algorithmic...
        greysphere 4y ago
        It takes work to create/identify/classify information, both in the economic and physics sense. That work should be allowed the same protections we do other forms of work. Your example is one where nearly no work was done, thus it doesn't deserve much value. "Let a = the set of all songs" doesn't help me find new songs I like. A songwriter does that work. Another artist that takes and uses and resells that work (without consent), is stealing that work. To me it's funny that nearly all the problem with the current team of AI generation would be solved if the model generators simply licensed the content they train on. "But that would cost too much" Ok, just use public domain work, "But that wouldn't be as good" Oh so you are saying the work has value, but you are unwilling to pay for it, and instead your scheme is to just take it. That seems like a good definition of stealing - not paying for something that has value.
rubinelli 4y ago
De minimis is a longstanding defense in copyright law. If you are copying very little from very many works, as is the case when you turn multiple petabytes into a few gigabytes of neural network weights, you are in the clear. The problem arises when models overfit and spit out almost perfect copies of the training data.
- phoe-krk 4y ago
  Yes, agreed, I don't think the problem is with networks that mix tons of input data in a way that doesn't heavily derive from one or a couple of sources. The currently available models do not have overfitting solved, though, and this technological imperfection also has direct practical (and legal) consequences.
- belorn 4y ago
  Copyright doesn't have an explicit size, but rather uses size as one of many indicators. For example, I could take a massive 8k video and covert it into a very small 144p youtube video. Am I in the clear simply because the output is tiny compared to the input? Similar I could take a huge studio master copy of a song and convert it to a very small and rather compressed (distorted) mp3. I partially agree that some of the problem is when perfect copies are spit out by the models, but I do think there is a bigger problem. Copyright is a complex concept that can't be defined exclusively by a single metric like size, and any mathematically definition will in the end be killed if large copyright holders feel threatened by it.
  - sdenton4 4y ago
    Thumbnail images don't violate copyright, and are a very helpful comparison case to consider. "Transformative Use" is a major consideration in fair use copyright: https://en.wikipedia.org/wiki/Transformative_use https://en.wikipedia.org/wiki/Transformative_use ML models do not supplant the pre-existing work, and provide fundamentally new modalities. Transformative use seems like a slam dunk to me, but I guess we'll see what the Supremes decide in twenty years or so...
    - 4y ago
yazzku 4y ago
I found all this to be very odd take coming from something entitled "copyleft currents". You summarized some of the points quite well. > So, all neural network developers, get ready for the lawyers, because they are coming to get you. This is just dumb.
CuriouslyC 4y ago
Incorrect. Training images are used to generate a latent manifold which might not contain any of the images in the original training set to within a meaningful delta unless they're massively overrepresented or cliche.
- phoe-krk 4y ago
  This might be technically correct but doesn't seem to matter much in practice because of how many practical cases of NN outputting copyright-infriging content there are. These include examples in the recent lawsuits that made the rounds on HN. Either even the best specialists behind these NNs cannot make the NNs not contain data that is "massively overrepresented or cliche", or they are unwilling to.
  - CuriouslyC 4y ago
    Or they're focusing on performance to get the tools to the place where they're good enough to start being adopted. Once getting sued is more of a problem than not having a product at all, it's not hard to switch gears. I imagine there are a number of ways to avoid storing "too close" copies in a model that have various tradeoffs, I'm sure they'll be adopted quickly in the face of litigation.
  - fzil 4y ago
    > the recent lawsuits that made the rounds on HN I happened to have missed those discussions, do you have some links you can point towards? thanks!
    - phoe-krk 4y ago
      https://stablediffusionlitigation.com/ https://stablediffusionlitigation.com/ https://githubcopilotlitigation.com/ https://githubcopilotlitigation.com/ Both along with their respective HN threads.
      - fzil 4y ago
        thanks a lot!
mdp2021 4y ago
> But they are compressed lossy copies of all that data So are those you are storing in your mind... (Edit: i.e., "learning is not a violation". Also see below.)
- scarmig 4y ago
  Well, if you copied a famous piece of art from memory... That'd be a copyright violation.
  - mdp2021 4y ago
    I meant that _learning_ is not a copyright violation. I do not see much space for misunderstanding: which relevant box takes instances of input data to output one of them? Edit: Make your point explicit, sniper... You can hide in front of the ### consuetude of "silent disagreement", but it remains violently annoying. The article contains "nuances" like "derivative work", but the poster is objecting to the article calling «neural networks» not «copies», as «compressed lossy copies», and I retorted that so is everything you learn, and holding them in the "corpus" is not considered "retaining a copy". If you have objections to that, either you present them, or there is no contribution in shaking your invisible head.
  - Groxx 4y ago
    Yeah, I think this is the crux of the issue that people keep glossing over when claiming it's not a true reproduction. Drawing Coca Cola's logo from memory by hand and slapping it on your product is still copyright infringement, even if it isn't an image Coca Cola has ever produced. In that sense it doesn't matter at all if it's AI or human - the production and subsequent distribution for profit of a copyrighted thing is not allowed, period. The current set of AIs do exactly this all the time. That's a very clear legal problem.
    - sdenton4 4y ago
      well, you get five demerits as the coca cola logo is a trademark, covered by a whole other branch of IP law...
tshaddox 4y ago
> But they are compressed lossy copies of all that data! Sounds similar to, for instance, songs written by people who have heard other songs. I wouldn’t expect legal cases concerning AI-generated works to be any simpler than legal cases concerning the difference between a songwriter violating the IP of another songwriter or simply being inspired by another song.
- lelanthran 4y ago
  Sometimes a songwriter violates copyright on their own song, without realising it.

tomrod 4y ago

Transgressors, perhaps, but not criminals. AI, especially generative AI, are not lossy copies but rather synthesizers. If anything, your comment highlights the need to evolve IP norms and laws.

thewakalix 4y ago

How are you using the word “copy”? It doesn’t seem to match the standard meaning. For instance, most people would not consider a brief summary of a movie’s plot to be a “copy” of that movie, or protected under copyright.

phoe-krk 4y ago
If you have an image, then train a neural network on that image, then use the neural network to reconstruct that image in detail, then the NN by definition contains enough information used to reconstruct that image - hence, a copy. With NNs trained on thousands or millions of data entries, this concept becomes fuzzy in the same way as you described - a short summary likely wouldn't be considered a copy, just like a 64x64 generated thumbnail wouldn't be considered in the same way a 4096x4096 hi-res image.
- Krasnol 4y ago
  > If you have an image, then train a neural network on that image, then use the neural network to reconstruct that image in detail I haven't seen that happening since the discussion started. Most of the complains I saw aimed at things like "it stole my style" not "it reproduced my art". Do you have any examples?
  - greysphere 4y ago
    It's more, 'this product is profiting from my labor without my consent (ie paying me).' In music you aren't allowed to use the same notes, even if you played them on a trumpet with a swing beat, while the source was on the piano very staccato. While we don't have the same vocabulary for art, it's not unreasonable to expect similar protections.
    - wvenable 4y ago
      What you describe for music is already outrageous -- why do you think that needs to be extended to everything? https://www.vice.com/en/article/wxepzw/musicians-algorithmically-generate-every-possible-melody-release-them-to-public-domain https://www.vice.com/en/article/wxepzw/musicians-algorithmic...
      - greysphere 4y ago
        It takes work to create/identify/classify information, both in the economic and physics sense. That work should be allowed the same protections we do other forms of work. Your example is one where nearly no work was done, thus it doesn't deserve much value. "Let a = the set of all songs" doesn't help me find new songs I like. A songwriter does that work. Another artist that takes and uses and resells that work (without consent), is stealing that work. To me it's funny that nearly all the problem with the current team of AI generation would be solved if the model generators simply licensed the content they train on. "But that would cost too much" Ok, just use public domain work, "But that wouldn't be as good" Oh so you are saying the work has value, but you are unwilling to pay for it, and instead your scheme is to just take it. That seems like a good definition of stealing - not paying for something that has value.

rubinelli 4y ago

De minimis is a longstanding defense in copyright law. If you are copying very little from very many works, as is the case when you turn multiple petabytes into a few gigabytes of neural network weights, you are in the clear. The problem arises when models overfit and spit out almost perfect copies of the training data.

phoe-krk 4y ago
Yes, agreed, I don't think the problem is with networks that mix tons of input data in a way that doesn't heavily derive from one or a couple of sources. The currently available models do not have overfitting solved, though, and this technological imperfection also has direct practical (and legal) consequences.
belorn 4y ago
Copyright doesn't have an explicit size, but rather uses size as one of many indicators. For example, I could take a massive 8k video and covert it into a very small 144p youtube video. Am I in the clear simply because the output is tiny compared to the input? Similar I could take a huge studio master copy of a song and convert it to a very small and rather compressed (distorted) mp3. I partially agree that some of the problem is when perfect copies are spit out by the models, but I do think there is a bigger problem. Copyright is a complex concept that can't be defined exclusively by a single metric like size, and any mathematically definition will in the end be killed if large copyright holders feel threatened by it.
- sdenton4 4y ago
  Thumbnail images don't violate copyright, and are a very helpful comparison case to consider. "Transformative Use" is a major consideration in fair use copyright: https://en.wikipedia.org/wiki/Transformative_use https://en.wikipedia.org/wiki/Transformative_use ML models do not supplant the pre-existing work, and provide fundamentally new modalities. Transformative use seems like a slam dunk to me, but I guess we'll see what the Supremes decide in twenty years or so...
  - 4y ago

yazzku 4y ago

I found all this to be very odd take coming from something entitled "copyleft currents". You summarized some of the points quite well. > So, all neural network developers, get ready for the lawyers, because they are coming to get you. This is just dumb.

CuriouslyC 4y ago

Incorrect. Training images are used to generate a latent manifold which might not contain any of the images in the original training set to within a meaningful delta unless they're massively overrepresented or cliche.

phoe-krk 4y ago
This might be technically correct but doesn't seem to matter much in practice because of how many practical cases of NN outputting copyright-infriging content there are. These include examples in the recent lawsuits that made the rounds on HN. Either even the best specialists behind these NNs cannot make the NNs not contain data that is "massively overrepresented or cliche", or they are unwilling to.
- CuriouslyC 4y ago
  Or they're focusing on performance to get the tools to the place where they're good enough to start being adopted. Once getting sued is more of a problem than not having a product at all, it's not hard to switch gears. I imagine there are a number of ways to avoid storing "too close" copies in a model that have various tradeoffs, I'm sure they'll be adopted quickly in the face of litigation.
- fzil 4y ago
  > the recent lawsuits that made the rounds on HN I happened to have missed those discussions, do you have some links you can point towards? thanks!
  - phoe-krk 4y ago
    https://stablediffusionlitigation.com/ https://stablediffusionlitigation.com/ https://githubcopilotlitigation.com/ https://githubcopilotlitigation.com/ Both along with their respective HN threads.
    - fzil 4y ago
      thanks a lot!

mdp2021 4y ago

> But they are compressed lossy copies of all that data So are those you are storing in your mind... (Edit: i.e., "learning is not a violation". Also see below.)

scarmig 4y ago
Well, if you copied a famous piece of art from memory... That'd be a copyright violation.
- mdp2021 4y ago
  I meant that _learning_ is not a copyright violation. I do not see much space for misunderstanding: which relevant box takes instances of input data to output one of them? Edit: Make your point explicit, sniper... You can hide in front of the ### consuetude of "silent disagreement", but it remains violently annoying. The article contains "nuances" like "derivative work", but the poster is objecting to the article calling «neural networks» not «copies», as «compressed lossy copies», and I retorted that so is everything you learn, and holding them in the "corpus" is not considered "retaining a copy". If you have objections to that, either you present them, or there is no contribution in shaking your invisible head.
- Groxx 4y ago
  Yeah, I think this is the crux of the issue that people keep glossing over when claiming it's not a true reproduction. Drawing Coca Cola's logo from memory by hand and slapping it on your product is still copyright infringement, even if it isn't an image Coca Cola has ever produced. In that sense it doesn't matter at all if it's AI or human - the production and subsequent distribution for profit of a copyrighted thing is not allowed, period. The current set of AIs do exactly this all the time. That's a very clear legal problem.
  - sdenton4 4y ago
    well, you get five demerits as the coca cola logo is a trademark, covered by a whole other branch of IP law...

tshaddox 4y ago

> But they are compressed lossy copies of all that data! Sounds similar to, for instance, songs written by people who have heard other songs. I wouldn’t expect legal cases concerning AI-generated works to be any simpler than legal cases concerning the difference between a songwriter violating the IP of another songwriter or simply being inspired by another song.

lelanthran 4y ago
Sometimes a songwriter violates copyright on their own song, without realising it.