5 ms·

With your definition you can kiss spam filters, search engines, and recommendation engines goodbye. Here is a very simple case. A spammer sends you text that h

by tensor 3y ago

With your definition you can kiss spam filters, search engines, and recommendation engines goodbye. Here is a very simple case. A spammer sends you text that has attached a license saying you may not use this text for any purpose other than to be read by a human. You flag this as spam and your spam filter updates its model weights using this text. Suddenly you are not allowed to do anything with your spam filter model weights other than read the weights yourself.

lelanthran 3y ago
They won't go away, they'll just be open source. (Before you object, my assertion without evidence is just as good as your assertion without evidence).
- tensor 3y ago
  I didn't state anything requiring evidence. I merely continued the logic proposed by the OP. You are fully capable of verifying or refuting the continuation without extra information.
eqvinox 3y ago
> A spammer sends you text that has attached a license You have not chosen to receive that text, and therefore you are not bound by that license. It may mathematically seem to not make a difference, but legally it does.
- tensor 3y ago
  That seems like a stretch. I mean if someone just put some source code in your mailbox you don't suddenly get the rights to use it however you want right?
  - eqvinox 3y ago
    > I mean if someone just put some source code in your mailbox you don't suddenly get the rights to use it however you want right? You're applying way too much logic to a legal problem. If you ask a lawyer or judge about this, their first question will be "what intent was that source code mailed under, and were you the intended recipient?" If someone mails you a bunch of source code by accident, and it's reasonably obvious to you that it was by accident (which it will frequently be, because who the f*ck mails source code around?), you may in fact be required to destroy it. On the other hand if someone mails you the same code and you have reason to believe their intention was to spread it out no strings attached, yeah, you get the rights to use it however you want… except if the sender didn't have the right to do that to begin with… P.S.: "no strings attached" is also something that is impossible in some jurisdictions, since what you're doing might be required to be a contract of some kind, and contracts require bidirectional considerations. But at this point you really need a lawyer to explain the actual situation… P.P.S.: this is like that joke about writing on a brick "by accepting this brick through your window, you indemnify the thrower against all possible charges or damage resulting from this brick" and then chucking the brick through some storefront window.
    - tensor 3y ago
      If we ask a lawyer they will probably cite precedence such as Authors Guild vs Google (https://towardsdatascience.com/the-most-important-supreme-court-decision-for-data-science-and-machine-learning-44cfc1c1bcaf https://towardsdatascience.com/the-most-important-supreme-co...) and thus this entire hypothetical interpretation of the legal framework is already not how the legal institutions see it. If the courts are to set new precedent I think it's important to consider all the downstream ramifications, and I think it's a lot more complex and challenging than a lot of people here seem to think. There is a lot more to AI than just generative neural networks. A lot of "boring" technology we all take for granted can be caught up in it.
      - eqvinox 3y ago
        > If we ask a lawyer they will probably cite precedence […] interpretation of the legal framework is already not how the legal institutions see it. Well, now this is an entirely different discussion, and FYI "precedence" is only a thing in half of the world's legal systems. Specifically, the common law (English) half. The other half, civil law (French) based systems, have no concept of "precedence"; verdicts from other courts have absolutely no law-like meaning. For every case and every judge, they are supposed to find the correct, applicable meaning of the laws as written by the legislative. Personally speaking, I find the "precedence" approach taken by English / Common Law incredibly silly — and actively harmful, it intermixes two branches of power (legislative & judicative) that should be 100% separated. Judges' interpretative rulings should not have (almost) the same effect as the legislative passing a law. (Google "Common Law vs Civil Law" for more info.) Anyway the original argument was that you could apply some license-like terms onto spam mails, and for that — no, you very much can't. The situation for AI is, to my knowledge, still very muddy at this point.
        tensor 3y ago
        I very much disagree that you can't specify license terms on spam. Lawyers certainly seem to think you can, as they always have a huge legal blob at the bottom of their emails about what you can or can't do with their email.
cornholio 3y ago
I agree the definition is overly simplified. The main point was that training implicitly contaminates the model with a presumption of derivation, training is not some magic pixie dust you can sprinkle onto protected works and strip copyright away. The next thing to discuss is if the derivative work is sufficiently transformative to be considered fair use without legal authorization from the rights owner. There is no easy analogy that can be made here, the type of derivation - transformation we talk about has no precedent in intellectual property law. My position to your challenge is that an AI text classifier/filter/recommendation engine is a sufficiently transformative derivation of the copyrighted works, whereas a general purpose machine that can produce works similar in style, content and character with the originals is not sufficiently transformative and should require authorization. The way I propose we arrive at that conclusion, without any legal precedent, is one based on first principles, on the intent and final purpose of copyright law. It's a political philosophy position, namely that intellectual property is a social convention designed around the creation of a common good, it exists to promote "the progress of science and useful arts", general flow of ideas and knowledge, by creating an economic incentive for creators to produce and make public their work - as opposed to keeping it secret to control its distribution, or abandoning creation for other fields, both net negative outcomes that diminish the public good. So when judging if the work of thinking machines is sufficiently transformative, we should ask: is what they output a net positive contribution to the common good of creation and widely available intellectual works, ideas or art? Is it at least not-negative? It's easy to make that argument with text classifiers, but much harder for something like Stable Diffusion. The algorithm it runs is completely dependent on human produced artworks, it cannot function without such an input and can't even produce a single creative bush stroke without them. Yet, the works it remixes using creative features of the originals can and indeed have already started to economically replace the work of original artists in the market place. So treating that derivation as fair use pushes society into a bad equilibrium, where artworks are less valuable and less likely to be produced, while the AI machine owner appropriates much of the economic value of the works it slurped in training. That's not 'fair use', and the AI machine as a whole is not 'fair use', even if some, or even all, of the works it produces could be considered taken individually, as fair use of the originals. This will continue to be true for as long as human creators remain a key ingredient of the automated creation process. When and if a machine can start to paint after reading an university course on painting, then that would be indeed a fair use of those manuals.

by tensor 3y ago

lelanthran 3y ago

They won't go away, they'll just be open source. (Before you object, my assertion without evidence is just as good as your assertion without evidence).

tensor 3y ago
I didn't state anything requiring evidence. I merely continued the logic proposed by the OP. You are fully capable of verifying or refuting the continuation without extra information.

eqvinox 3y ago

> A spammer sends you text that has attached a license You have not chosen to receive that text, and therefore you are not bound by that license. It may mathematically seem to not make a difference, but legally it does.

tensor 3y ago
That seems like a stretch. I mean if someone just put some source code in your mailbox you don't suddenly get the rights to use it however you want right?
- eqvinox 3y ago
  > I mean if someone just put some source code in your mailbox you don't suddenly get the rights to use it however you want right? You're applying way too much logic to a legal problem. If you ask a lawyer or judge about this, their first question will be "what intent was that source code mailed under, and were you the intended recipient?" If someone mails you a bunch of source code by accident, and it's reasonably obvious to you that it was by accident (which it will frequently be, because who the f*ck mails source code around?), you may in fact be required to destroy it. On the other hand if someone mails you the same code and you have reason to believe their intention was to spread it out no strings attached, yeah, you get the rights to use it however you want… except if the sender didn't have the right to do that to begin with… P.S.: "no strings attached" is also something that is impossible in some jurisdictions, since what you're doing might be required to be a contract of some kind, and contracts require bidirectional considerations. But at this point you really need a lawyer to explain the actual situation… P.P.S.: this is like that joke about writing on a brick "by accepting this brick through your window, you indemnify the thrower against all possible charges or damage resulting from this brick" and then chucking the brick through some storefront window.
  - tensor 3y ago
    If we ask a lawyer they will probably cite precedence such as Authors Guild vs Google (https://towardsdatascience.com/the-most-important-supreme-court-decision-for-data-science-and-machine-learning-44cfc1c1bcaf https://towardsdatascience.com/the-most-important-supreme-co...) and thus this entire hypothetical interpretation of the legal framework is already not how the legal institutions see it. If the courts are to set new precedent I think it's important to consider all the downstream ramifications, and I think it's a lot more complex and challenging than a lot of people here seem to think. There is a lot more to AI than just generative neural networks. A lot of "boring" technology we all take for granted can be caught up in it.
    - eqvinox 3y ago
      > If we ask a lawyer they will probably cite precedence […] interpretation of the legal framework is already not how the legal institutions see it. Well, now this is an entirely different discussion, and FYI "precedence" is only a thing in half of the world's legal systems. Specifically, the common law (English) half. The other half, civil law (French) based systems, have no concept of "precedence"; verdicts from other courts have absolutely no law-like meaning. For every case and every judge, they are supposed to find the correct, applicable meaning of the laws as written by the legislative. Personally speaking, I find the "precedence" approach taken by English / Common Law incredibly silly — and actively harmful, it intermixes two branches of power (legislative & judicative) that should be 100% separated. Judges' interpretative rulings should not have (almost) the same effect as the legislative passing a law. (Google "Common Law vs Civil Law" for more info.) Anyway the original argument was that you could apply some license-like terms onto spam mails, and for that — no, you very much can't. The situation for AI is, to my knowledge, still very muddy at this point.
      - tensor 3y ago
        I very much disagree that you can't specify license terms on spam. Lawyers certainly seem to think you can, as they always have a huge legal blob at the bottom of their emails about what you can or can't do with their email.

cornholio 3y ago

I agree the definition is overly simplified. The main point was that training implicitly contaminates the model with a presumption of derivation, training is not some magic pixie dust you can sprinkle onto protected works and strip copyright away. The next thing to discuss is if the derivative work is sufficiently transformative to be considered fair use without legal authorization from the rights owner. There is no easy analogy that can be made here, the type of derivation - transformation we talk about has no precedent in intellectual property law. My position to your challenge is that an AI text classifier/filter/recommendation engine is a sufficiently transformative derivation of the copyrighted works, whereas a general purpose machine that can produce works similar in style, content and character with the originals is not sufficiently transformative and should require authorization. The way I propose we arrive at that conclusion, without any legal precedent, is one based on first principles, on the intent and final purpose of copyright law. It's a political philosophy position, namely that intellectual property is a social convention designed around the creation of a common good, it exists to promote "the progress of science and useful arts", general flow of ideas and knowledge, by creating an economic incentive for creators to produce and make public their work - as opposed to keeping it secret to control its distribution, or abandoning creation for other fields, both net negative outcomes that diminish the public good. So when judging if the work of thinking machines is sufficiently transformative, we should ask: is what they output a net positive contribution to the common good of creation and widely available intellectual works, ideas or art? Is it at least not-negative? It's easy to make that argument with text classifiers, but much harder for something like Stable Diffusion. The algorithm it runs is completely dependent on human produced artworks, it cannot function without such an input and can't even produce a single creative bush stroke without them. Yet, the works it remixes using creative features of the originals can and indeed have already started to economically replace the work of original artists in the market place. So treating that derivation as fair use pushes society into a bad equilibrium, where artworks are less valuable and less likely to be produced, while the AI machine owner appropriates much of the economic value of the works it slurped in training. That's not 'fair use', and the AI machine as a whole is not 'fair use', even if some, or even all, of the works it produces could be considered taken individually, as fair use of the originals. This will continue to be true for as long as human creators remain a key ingredient of the automated creation process. When and if a machine can start to paint after reading an university course on painting, then that would be indeed a fair use of those manuals.