OpenAI now tries to hide that ChatGPT was trained on copyrighted books, including J.K. Rowling’s Harry Potter series::A new research paper laid out ways in which AI developers should try and avoid showing LLMs have been trained on copyrighted material.

  • BURN@lemmy.world
    link
    fedilink
    English
    arrow-up
    1
    arrow-down
    1
    ·
    1 year ago

    Nowhere did I say derivative works are not ok. If a game maker explicitly forbids using modded versions of their game, I think that should be up to them. Games that have vibrant modding communities are almost always at least partially supported by the developer anyways.

    My points are individual copyright anyways, not corporate. With increasing individual protections I also propose decreasing corporate copyright protection.

    I believe that China makes 90% of the same product for 80% of the price after ripping off their American counterparts. But that’s also entirely off topic and really has nothing to do with this. Art/Creative Works are entirely different than physical goods.

    • THE_STORM_BLADE@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 year ago

      How is what AI produces not derivative? Like humans, AI takes in a bunch of inputs (think about all the art you’ve seen, read, and watched, and how it affects the art you create), and outputs something that’s derivative from the input.

      • BURN@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        arrow-down
        1
        ·
        1 year ago

        Because AI has nothing new to add to the work. It’s only able to use work that it’s even before to add, and can’t learn from anything being created.

        There’s no new generation. AI does not work like a human and should not be afforded the same rights as a person. AI does not transform works the way a human would.