He allegedly used Stable Diffusion, a text-to-image generative AI model, to create “thousands of realistic images of prepubescent minors,” prosecutors said.
  • @PotatoKat@lemmy.worldEnglish
    -4•7 months ago

    Real children are in training data regardless of if there is csam in the data or not (which there is a high chance there is considering how they get their training data) so real children are involved

    • @Duamerthrax@lemmy.worldEnglish
      4•7 months ago

      I’ve already stated that I do not support using images of real children in the models. Even if the images are safe/legal, it’s a violation of privacy.