3 ms·

AI doesn't get inspired. It's not human. It adds everything about it to its endless stream of levers to pull, and if you pull the right ones, it will just give

by tapoxi 3y ago

AI doesn't get inspired. It's not human. It adds everything about it to its endless stream of levers to pull, and if you pull the right ones, it will just give you the source verbatim as proven by the NYT lawsuit filing where it was just outputting unaltered copywritten NYT article text.

PretzelPirate 3y ago
> as proven by the NYT lawsuit filing I would be careful in asserting that it's proven until the evidence is heard in court and we have all the data.
- kevingadd 3y ago
  You can look at the filing yourself. It is unambiguous.
  - ShamelessC 3y ago
    It’s unambiguous about repeating articles of NY times pieces with hallucinations interwoven occasionally. This isn’t nearly as strong of an argument.
ta8645 3y ago
That's a matter of perspective. AI's do not make a copy of the source material. It very much just adjusts their internal weights, which from a broadminded perspective, can be seen as simple inspiration, and not copying. Of course, just like a human artist, it could probably closely approximate the source material if it wanted to, but it would still be its own approximation, not an exact duplicate. As for plagiarism, we all have to be careful to rephrase things we've read elsewhere, and perhaps AI's need to be trained to do this a bit better... but it doesn't change the underlying fact that they're learning from the text they read, not storing a verbatim copy (at least, no more so than a human reader with a good memory)
- ImPostingOnHN 3y ago
  Based on the comment you replied to, it seems they are indeed producing verbatim copies.
  - ta8645 3y ago
    > Based on the comment you replied to, it seems they are indeed producing verbatim copies. Well i'll leave it to the legal system to decide if that's true. But in any case, that's no different from a human with a photographic memory doing the same thing after reading a paragraph. We don't blame them for their superior memory, or being inspired by the knowledge. We don't claim they've violated copyright because their memory contains an exact copy of what they read. We may still demand that they avoid reproducing the exact words they've read, even though they are capable of it -- which is fine. We can demand the same of AI's. All I object to, is the idea that a smart AI, with a great memory is guilty of something just by reading or viewing content that was willingingly shared online.
    - kevingadd 3y ago
      If I tell you water is wet and the sky is blue, will you be waiting for a court case to grind through the appeals process on that as well? The examples in the filing were unambiguous. You can go look it up and see them, they were also cited in all the news articles I saw about it. The AI regurgitated many paragraphs of text with very, very few small modifications.
      - ta8645 3y ago
        The issue at hand is not if some words were copied; it's a legal issue of whether that constitutes legal or otherwise fair use or not. And I'm not a lawyer, and am happy to wait for the court to decide. But to be honest, I don't really care one way or the other, since it doesn't get to the heart of the matter as I see it. To my mind, it's no different than a human with a good memory doing the same thing. Specifically, it isn't the consumption of the media that is the problem, or even remembering it very well. Rather, it is the public verbatim reproduction of it, which a human might inappropriately do as well. AI's need to be trained to avoid unfair use of source material, but I don't think that means they should be prohibited from consuming public material.
CrimsonRain 3y ago
If you pull the right levers, you can also copy the NYT article fully.
- tapoxi 3y ago
  Yeah and that's copyright infringement. That's why if you're reverse engineering something it needs to be done in a clean room environment, your prior exposure to the copywritten material poisons the well for any derivative you create. This extends to music as well, if someone hears a song and is inspired by that in their work, the original artist gets credit.
  - ta8645 3y ago
    > This extends to music as well, if someone hears a song and is inspired by that in their work, the original artist gets credit. That's only true in a very narrow set of circumstances. Imagine the case where someone listens to 10,000 songs, and then takes the sum total of that experience and writes their own. There's no credit given to the inspiration that each of those 10,000 songs gave. And that is in fact much closer to what AI is currently doing.
    - tapoxi 3y ago
      If we're discussing the current situation, where ChatGPT is outputting entire articles of NYT copyrighted content, then it certainly matches say, Bittersweet Symphony containing a small sample of a Rolling Stones song resulting in the Stones getting credit for the entire work.
      - ta8645 3y ago
        There is a rather vibrant culture of _human_ remix artists who sample existing music and generate something completely new. It is still an open question of how long a sample they can use before it requires licensing: https://www.ajschwartzlaw.com/can-i-use-that-part-2-remix-culture-the-internet-and-copyright-law/ https://www.ajschwartzlaw.com/can-i-use-that-part-2-remix-cu... But all of this, doesn't have much to do with whether an AI should be allowed to consume public media -- even if we agree they need to do a better job about avoiding verbatim reconstruction when prompted. Edit: I wasn't able to find the Lawrence Lessig talk I wanted to link, but here's a quite old one that makes some of the same points: https://youtu.be/X8ULxxgjBuI https://youtu.be/X8ULxxgjBuI
imgabe 3y ago
...and if you ask a human artist to exactly reproduce an artwork, that is copyright infringment, or forgery. I don't see how this is any different? The person using the AI to make exact replicas would be the one committing the copyright infringement, not the AI.

by tapoxi 3y ago

PretzelPirate 3y ago

> as proven by the NYT lawsuit filing I would be careful in asserting that it's proven until the evidence is heard in court and we have all the data.

kevingadd 3y ago
You can look at the filing yourself. It is unambiguous.
- ShamelessC 3y ago
  It’s unambiguous about repeating articles of NY times pieces with hallucinations interwoven occasionally. This isn’t nearly as strong of an argument.

ta8645 3y ago

That's a matter of perspective. AI's do not make a copy of the source material. It very much just adjusts their internal weights, which from a broadminded perspective, can be seen as simple inspiration, and not copying. Of course, just like a human artist, it could probably closely approximate the source material if it wanted to, but it would still be its own approximation, not an exact duplicate. As for plagiarism, we all have to be careful to rephrase things we've read elsewhere, and perhaps AI's need to be trained to do this a bit better... but it doesn't change the underlying fact that they're learning from the text they read, not storing a verbatim copy (at least, no more so than a human reader with a good memory)

ImPostingOnHN 3y ago
Based on the comment you replied to, it seems they are indeed producing verbatim copies.
- ta8645 3y ago
  > Based on the comment you replied to, it seems they are indeed producing verbatim copies. Well i'll leave it to the legal system to decide if that's true. But in any case, that's no different from a human with a photographic memory doing the same thing after reading a paragraph. We don't blame them for their superior memory, or being inspired by the knowledge. We don't claim they've violated copyright because their memory contains an exact copy of what they read. We may still demand that they avoid reproducing the exact words they've read, even though they are capable of it -- which is fine. We can demand the same of AI's. All I object to, is the idea that a smart AI, with a great memory is guilty of something just by reading or viewing content that was willingingly shared online.
  - kevingadd 3y ago
    If I tell you water is wet and the sky is blue, will you be waiting for a court case to grind through the appeals process on that as well? The examples in the filing were unambiguous. You can go look it up and see them, they were also cited in all the news articles I saw about it. The AI regurgitated many paragraphs of text with very, very few small modifications.
    - ta8645 3y ago
      The issue at hand is not if some words were copied; it's a legal issue of whether that constitutes legal or otherwise fair use or not. And I'm not a lawyer, and am happy to wait for the court to decide. But to be honest, I don't really care one way or the other, since it doesn't get to the heart of the matter as I see it. To my mind, it's no different than a human with a good memory doing the same thing. Specifically, it isn't the consumption of the media that is the problem, or even remembering it very well. Rather, it is the public verbatim reproduction of it, which a human might inappropriately do as well. AI's need to be trained to avoid unfair use of source material, but I don't think that means they should be prohibited from consuming public material.

CrimsonRain 3y ago

If you pull the right levers, you can also copy the NYT article fully.

tapoxi 3y ago
Yeah and that's copyright infringement. That's why if you're reverse engineering something it needs to be done in a clean room environment, your prior exposure to the copywritten material poisons the well for any derivative you create. This extends to music as well, if someone hears a song and is inspired by that in their work, the original artist gets credit.
- ta8645 3y ago
  > This extends to music as well, if someone hears a song and is inspired by that in their work, the original artist gets credit. That's only true in a very narrow set of circumstances. Imagine the case where someone listens to 10,000 songs, and then takes the sum total of that experience and writes their own. There's no credit given to the inspiration that each of those 10,000 songs gave. And that is in fact much closer to what AI is currently doing.
  - tapoxi 3y ago
    If we're discussing the current situation, where ChatGPT is outputting entire articles of NYT copyrighted content, then it certainly matches say, Bittersweet Symphony containing a small sample of a Rolling Stones song resulting in the Stones getting credit for the entire work.
    - ta8645 3y ago
      There is a rather vibrant culture of _human_ remix artists who sample existing music and generate something completely new. It is still an open question of how long a sample they can use before it requires licensing: https://www.ajschwartzlaw.com/can-i-use-that-part-2-remix-culture-the-internet-and-copyright-law/ https://www.ajschwartzlaw.com/can-i-use-that-part-2-remix-cu... But all of this, doesn't have much to do with whether an AI should be allowed to consume public media -- even if we agree they need to do a better job about avoiding verbatim reconstruction when prompted. Edit: I wasn't able to find the Lawrence Lessig talk I wanted to link, but here's a quite old one that makes some of the same points: https://youtu.be/X8ULxxgjBuI https://youtu.be/X8ULxxgjBuI

imgabe 3y ago

...and if you ask a human artist to exactly reproduce an artwork, that is copyright infringment, or forgery. I don't see how this is any different? The person using the AI to make exact replicas would be the one committing the copyright infringement, not the AI.