• I_Has_A_Hat@lemmy.world
    link
    fedilink
    English
    arrow-up
    16
    ·
    edit-2
    1 day ago

    That’s not how AI works. The people claiming that AI models have to be trained on CP to make CP have a fundamental misunderstanding of how image generation models work.

    If I ask an AI model to make an image of a sexy air fryer fucking a sexy, anthropomorphic, ear of corn, it will produce said image. I promise it was not trained on sexy images of air fryers fucking sexy anthropomorphic ears of corn.

    That’s what makes AI so dangerous and difficult to regulate. We are not ready for this technology to be in the hands of everyone.

    • jpreston2005@lemmy.world
      link
      fedilink
      English
      arrow-up
      2
      ·
      1 day ago

      You’re wrong. Literally from the article

      AI-generated CSAM may also present other challenges not raised by the government in this appeal. For example, researchers have found that “a major dataset used for training image-generating AI contained hundreds of CSAM images.” Thus, unlike the virtual CSAM at issue in Free Speech Coalition, the market for AI-generated virtual CSAM images may impact the demand for actual CSAM on which the AI models can train.

      • Amnesigenic@lemmy.ml
        link
        fedilink
        English
        arrow-up
        1
        ·
        13 hours ago

        So get rid of that dataset, you’ll still have systems that can generate similar images anyway. If it can generate images of things that do not and cannot exist in reality then it can generate CSAM without using real CSAM.

    • _stranger_@lemmy.world
      link
      fedilink
      English
      arrow-up
      3
      ·
      1 day ago

      They’re not wrong about the ruling incentivising CP though. I’m sure we’ll see shitloads of sites outright advertising it now