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.

    • OkToBeTakei@lemm.ee
      link
      fedilink
      English
      arrow-up
      16
      arrow-down
      3
      ·
      edit-2
      1 year ago

      If you then use your brain to reproduce a product based on copyrighted work owned by Warner (without their permission) and try to sell it, then you could be in violation of the law.

      the thing is: AI and your brain are very different things

        • OkToBeTakei@lemm.ee
          link
          fedilink
          English
          arrow-up
          5
          arrow-down
          1
          ·
          1 year ago

          I suppose you could say I’m making a good-faith assumption. It’s possible that he is.

          my point is still valid.

          • kmkz_ninja@lemmy.world
            link
            fedilink
            English
            arrow-up
            5
            arrow-down
            3
            ·
            1 year ago

            His point is equally valid. Can an artist be compelled to show the methods of their art? Is it as right to force an artist to give up methods if another artist thinks they are using AI to derive copyrighted work? Haven’t we already seen that LLMs are really poor at evaluating whether or not something was created by an LLM? Wouldn’t making strong laws on such an already opaque and difficult-to-prove issue be more of a burden on smaller artists vs. large studios with lawyers-in-tow.

            • OkToBeTakei@lemm.ee
              link
              fedilink
              English
              arrow-up
              4
              arrow-down
              4
              ·
              edit-2
              1 year ago

              His point is equally valid.

              but it’s irrelevant.

              Can an artist be compelled to show the methods of their art? Is it as right to force an artist to give up methods if another artist thinks they are using AI to derive copyrighted work?

              none of this is relevant in copyright law. the only thing that matters is who published it first and who is then using that copyrighted work for profit without first having gotten permission of the owner.

              Haven’t we already seen that LLMs are really poor at evaluating whether or not something was created by an LLM?

              also irrelevant

              Wouldn’t making strong laws on such an already opaque and difficult-to-prove issue

              laws are not written with the idea of whether on they’re hard or easy to prove as a consideration. also, your claims of such things being easy/hard to prove is a matter of opinion, and I don’t agree with you chain of deduction here. OpenAI admitted that Rowlong’s works (among many others) were used for training chatGPT.

              be more of a burden on smaller artists vs. large studios with lawyers-in-tow.

              also irrelevant, unless you’re arguing that, because it’s too difficult for small artists to defend their copyrighted work, they should just shut up and deal with it. currently, there is no legal precedent as to whether this is or is not copyright infringement, which is what a lawsuit like this is intended to set. For most, I would say, it Eem that it clearly is, for others, it’s not so clear.

      • Asuka@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        4
        arrow-down
        1
        ·
        1 year ago

        If I read Harry Potter and wrote a novel of my own, no doubt ideas from it could consciously or subconsciously influence it and be incorporated into it. Hey is that any different from what an LLM does?

    • TropicalDingdong@lemmy.world
      link
      fedilink
      English
      arrow-up
      13
      arrow-down
      5
      ·
      1 year ago

      Exactly. If I write some Loony toons fan fiction, Warner doesn’t own that. This ridiculous view of copyright (that’s not being challenged in the public discourse) needs to be confronted.

      • OkToBeTakei@lemm.ee
        link
        fedilink
        English
        arrow-up
        12
        arrow-down
        4
        ·
        edit-2
        1 year ago

        that’s not exactly what’s in dispute— the prodcut that LLMs produce. That would probably be ruled as a derivative work under the DMCA’s “Fair Use” clause, and, therefore, public domain.

        the issue at hand is that the company accessed the copyrighted material without paying for it and is now using that training to earn more money without fair compensation.

        these language models or even proper AI can’t create original creative works the way a human can. The best it can do it create a pastiche or composition that simulates originality but is really just a jumble of recycled ideas that it’s been trained on. There’s a fair argument to be made that the owners of the copyrights of those pesos works are entitled to fair compensation, especially since AI is a tool used by a company to churn out profit off the work of others.

          • Sethayy@sh.itjust.works
            link
            fedilink
            English
            arrow-up
            1
            ·
            1 year ago

            Can’t but theyre pretty open on how they trained the model, so like almost admitted guilt (though they werent hosting the pirated content, its still out there and would be trained on). Cause unless they trained it on a paid Netflix account, there’s no way to get it legally.

            Idk where this lands legally, but I’d assume not in their favour

    • CoderKat@lemm.ee
      link
      fedilink
      English
      arrow-up
      6
      ·
      edit-2
      1 year ago

      It’s honestly a good question. It’s perfectly legal for you to memorize a copyrighted work. In some contexts, you can recite it, too (particularly the perilous fair use). And even if you don’t recite a copyrighted work directly, you are most certainly allowed to learn to write from reading copyrighted books, then try to come up with your own writing based off what you’ve read. You’ll probably try your best to avoid copying anyone, but you might still make mistakes, simply by forgetting that some idea isn’t your own.

      But can AI? If we want to view AI as basically an artificial brain, then shouldn’t it be able to do what humans can do? Though at the same time, it’s not actually a brain nor is it a human. Humans are pretty limited in what they can remember, whereas an AI could be virtually boundless.

      If we’re looking at intent, the AI companies certainly aren’t trying to recreate copyrighted works. They’ve actively tried to stop it as we can see. And LLMs don’t directly store the copyrighted works, either. They’re basically just storing super hard to understand sets of weights, which are a challenge even for experienced researchers to explain. They’re not denying that they read copyrighted works (like all of us do), but arguably they aren’t trying to write copyrighted works.

    • SubArcticTundra@lemmy.ml
      link
      fedilink
      English
      arrow-up
      2
      arrow-down
      2
      ·
      1 year ago

      No, because you paid for a single viewing of that content with your cinema ticket. And frankly, I think that the price of a cinema ticket (= a single viewing, which it was) should be what OpenAI should be made to pay.