• saltesc@lemmy.world
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 month ago

    Yep. Machines will only ever do what they’re told to do. This is AI literally doing the job it’s been instructed to do under the rules it has been given.

    • merc@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      2
      ·
      edit-2
      1 month ago

      Machines are not designed by hermits who have no knowledge of the outside world. They’re tools, but they’re tools designed with a purpose and with or without safeties designed to keep them from maiming or killing people. The design of the machine can be used to talk about the responsibility and morals of the machine’s designer. And, certain machines are so unsafe that even if they theoretically can have a useful purpose, the dangers of the machine being misused are so great that the machine shouldn’t be permitted to be sold.

      In Arrested Development, George Bluth designs and sells the Cornballer, a machine to deep-fry cornballs. It was made illegal after it caused serious burns to anybody who used it. Part of the purpose of showing this device on the show is to reveal the character of George Bluth. It shows that he’s the kind of guy who doesn’t care enough to design a safe device, and who continues to try to sell it in Mexico even after it’s made illegal in the US because of how unsafe it is.

      Yes, in this case it is people who are submitting papers full of fabricated data using ChatGPT as a tool. But, that doesn’t mean that ChatGPT is simply “neutral” in this whole thing. They’ve released a tool that lacks safeties and that is effectively “burning” science. The positive potential uses of ChatGPT are what, writing a dirty limerick in the style of Shakespeare? Meanwhile, the potential pitfalls of using it are things like having it convince a suicidal person to kill themselves, sowing confusion and making it harder to find good science, giving people unsafe medical diagnoses?