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.

    • FatCrab@lemmy.one
      link
      fedilink
      English
      arrow-up
      6
      arrow-down
      3
      ·
      1 year ago

      The reasoning that claims training a generative model is infringing IP would still mean a robot going into a library with a card it has to optically read all the books there to create the same generative model would still be infringing IP.

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

      Humans can judge information make decisions on it and adapt it. AI mostly just looks at what is statistically what is most likely based on training data. If 1 piece of data exists, it will copy, not paraphrase. Example was from I think copilot where it just printed out the code and comments from an old game verbatim. I think Quake2. It isn’t intelligence, it is statistical copying.

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

        Well, mathematics cannot be copyrighted. In most countries at least.