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.

  • Technoguyfication@lemmy.ml
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    People are acting like ChatGPT is storing the entire Harry Potter series in its neural net somewhere. It’s not storing or reproducing text in a 1:1 manner from the original material. Certain material, like very popular books, has likely been interpreted tens of thousands of times due to how many times it was reposted online (and therefore how many times it appeared in the training data).

    Just because it can recite certain passages almost perfectly doesn’t mean it’s redistributing copyrighted books. How many quotes do you know perfectly from books you’ve read before? I would guess quite a few. LLMs are doing the same thing, but on mega steroids with a nearly limitless capacity for information retention.

    • Teritz@feddit.de
      link
      fedilink
      English
      arrow-up
      0
      arrow-down
      1
      ·
      1 year ago

      Using Copyrighted Work as Art as example still influences the AI which their make Profit from.

      If they use my Works then they need to pay thats it.

      • coheedcollapse@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        ·
        1 year ago

        Still kinda blows my mind how like the most socialist people I know (fellow artists) turned super capitalist the second a tool showed like an inkling of potential to impact their bottom line.

        Personally, I’m happy to have my work scraped and permutated by systems that are open to the public. My biggest enemy isn’t the existence of software scraping an open internet, it’s the huge companies who see it as a way to cut us out of the picture.

        If we go all copyright crazy on the models for looking at stuff we’ve already posted openly on the internet, the only companies with access to the tools will be those who already control huge amounts of data.

        I mean, for real, it’s just mind-blowing seeing the entire artistic community pretty much go full-blown “Metallica with the RIAA” after decades of making the “you wouldn’t download a car” joke.

        • Sir_Kevin@lemmy.dbzer0.com
          link
          fedilink
          English
          arrow-up
          1
          ·
          1 year ago

          Fuckin preach! I feel like I’m surrounded by children that didn’t live through the many other technologies that have came along and changed things. People lost their shit when photoshop became mainstream, when music started using samples, etc. AI is here to stay. These same people are probably listening to autotuned music all day while they complain on the internet about AI looking at their art.

        • angstylittlecatboy@reddthat.com
          link
          fedilink
          English
          arrow-up
          1
          ·
          edit-2
          1 year ago

          I feel like a lot of internet people (not even just socialists) go from seeing copyright as at best a compromise that allows the arts to have value under capitalism to treating it like a holy doctrine when the subject of LLMs comes up.

          Like, people who will say “piracy is always okay” will also say “ban AI, period” (and misrepresent organizations that want regulations on it’s use as wanting a full ban.)

          Like, growing up with an internet full of technically illegal content (or grey area at best) like fangames and YouTube Poops made me a lifelong copyright skeptic. It’s outright confusing to me when people take copyright as seriously as this.