Article by IEEE Spectrum: The writers tested the two AI image generators MidJourney and Stable Diffusion, testing their abilities to generate imagery that closely resembled copyrighted material, which proves that the training data of the image generators had to contain copyrighted material. Implemented safeguards were largely unsuccessful to curb the output of potentially infringing images.

  • pixxelkick@lemmy.world
    link
    fedilink
    arrow-up
    9
    ·
    10 months ago

    Everytime this comes up I have to ask: so?

    Artists can, and have, done the same thing by hand with countless other tools before.

    The capability to or not to infringe copyrights with (tool) isn’t what has defined if people can get away with doing it.

    Plagiarism laws are largely reactionary law. You have to prove the person ripped off your IP and profited off it.

    People have been trying to make money off other people’s IP covertly for all eternity, midjourney isn’t very special.

    • squirrel@lemmy.blahaj.zoneOPM
      link
      fedilink
      English
      arrow-up
      13
      ·
      10 months ago

      Because it’s a difference in scale: Sure, there are a handful of artists out there who can emulate a style/create a picture to the point where it is indistinguishable from somebody else’s copyrighted work. But the AI companies offer the tools for everyone to do that and they monetize this work of theirs (as in selling access to AI tools), while telling everybody else that rights holders should not be able to monetize their work. It’s a blatantly obvious double standard.

      • AnonStoleMyPants@sopuli.xyz
        link
        fedilink
        arrow-up
        4
        ·
        10 months ago

        Can you explain the double standard a bit more? I don’t understand it. Are you saying that the double standard is that AI companies sell a product that can be used to infringe copyright, yet they say that people infringing it using this sold product cannot monetize it?

        • squirrel@lemmy.blahaj.zoneOPM
          link
          fedilink
          English
          arrow-up
          7
          ·
          10 months ago

          What I mean is this: AI companies are arguing that people should not be able to earn money from the works they created (for example through selling licenses to their copyrighted works), insofar as paying for their training data is concerned. While - at the same time - they are charging money for the creation of works with AI.

          To put it differently: “Artists should not earn money from creation of artworks. We should earn money from creation of artworks.”

    • Franzia@lemmy.blahaj.zone
      link
      fedilink
      arrow-up
      6
      ·
      10 months ago

      When artists copy someone else’s style or use an IP (I imagine Patreon fanfic art of trendy characters) both the original studio and the copying artist get paid for their labor.

      With AI only the AI company gets paid in subscriptions for the service of operating a server that has the AI algorithm running on it.

      My basic take is that I want artists paid for their work and the AI company, which has insane funding and legal responsibility, is stealing.

      If Disney sends an individual artist a cease and desist - that artist almost certainly complies. If Disney sent the AI company a cease and desist - its a legal battle designed to cement, not dismantle, copyright infringement rules - for the enrichment of those who already have structural power in the job market for art creation.

  • mindbleach@sh.itjust.works
    link
    fedilink
    arrow-up
    2
    ·
    10 months ago

    “The machine did what we asked!!!” is an increasingly common and consistently stupid headline.

    Absolutely nobody promised the draw-anything machine would invent perfectly unique images, unburdened by any prior intellectual property concerns. How the fuck would it. We fed it every image on the internet, with labels; no kidding it knows what Darth Vader looks like. And it’ll emit images of him in a Star Wars-y context, with other related elements. That’s what it’s for.

    Ask for a deer and you get a forest.

    Ask for trees and it should pick a particular kind of tree, rather than making up some whole new species that’s never been seen. Ask for images of a cartoon family, and yeah, sometimes that will be the Simpsons. What else was it supposed to show you? A lot of example trees are oak. A lot of example cartoons are the Bart.

    When you explicitly name a character and a movie, you have negative room to complain that the machine did its fucking job. Cutesy indirect descriptions are no better. You know damn well which cartoon families have yellow skin, and which sci-fi series have laser swords. Don’t type “tree with needles instead of leaves” and go “Ah-HA! Spruce!”

  • Ranger@lemmy.blahaj.zone
    link
    fedilink
    arrow-up
    1
    ·
    10 months ago

    Also this doesn’t prove that they used copyrighted materials because there are exceptions to copyright & gray areas like fan art.

  • Ranger@lemmy.blahaj.zone
    link
    fedilink
    arrow-up
    1
    ·
    edit-2
    10 months ago

    IA doesn’t copy/paste images, it creates internal concepts of images/labels & creates a random image & changes it until it’s in-line with it’s internal concepts. It’s not much different then if you asked an artist to draw something in someone’s else’s style, which Disney has done for years.