Link

AI have no rights. Your AI creations are right-less. They belong in the public domain. If not, they are properties of the peoples whose art you stole to make the AI.

  • bazingabrain [comrade/them]
    ·
    edit-2
    7 months ago

    Yes it makes image generation super easy and accessible to people without technical skills but so did Photoshop so did the camera so did fucking crayons.

    moronic take michael-laugh

      • GarbageShoot [he/him]
        ·
        7 months ago

        Legally, you can't take a photo of someone else's painting and just declare it your own, claiming they were an artist who made the image with one set of tools and you are another artist who made the image with "just another set of tools". Making a collage of different paintings that fails to be a transformative use of any of them also isn't protected. This is a derivative of that.

        • Commiejones [comrade/them, he/him]
          ·
          7 months ago

          Legally, you can't take a photo of someone else's painting and just declare it your own, claiming they were an artist who made the image with one set of tools

          You cant use AI to remake another artwork and claim it is your own either. Plagiarism isn't copyrightable no matter what tools you use.

          AI art is less derivative than most art. Usually it is based on thousands to millions of other images if that isn't transformative than what is?

          • GarbageShoot [he/him]
            ·
            7 months ago

            First, you cut out the more relevant example, second, you don't know what transformative means in this context.

            Effectively, a transformative use is one where the media (or whatever) is used for a purpose that is very different from whatever the original purpose is, e.g. featuring a painting in the background of a comic or movie to add some kind of thematic coding to a scene.

            What an AI uses training data for is literally the opposite. It uses paintings of cars to mathematically establish how to produce an image that looks like a painting of a car. It is very specifically using the data in order to accomplish exactly what it thinks the samples are accomplishing. If it does not view information as being pertinent to car paintings, it does not use it for making a car painting.