• @[email protected]
    link
    fedilink
    English
    643 months ago

    Okay now let’s ask it to write anti ransomware. My guess is it will help with that too. And then the balance is struck and the obvious becomes obvious: AI is a tool to enhance all aspects of our lives. But instead we seem to only hear about the ways we should be fearful and worried about it.

    • SeaJ
      link
      fedilink
      English
      173 months ago

      We should just ask it to continually play tic tac toe against itself and finish it off for good.

      • Decoy321
        link
        fedilink
        English
        33 months ago

        After all, the only winning move is not to play.

    • @[email protected]
      link
      fedilink
      English
      93 months ago

      Well then I’ll just ask it to write a program to make a program to … Un-enhance … De-enhance?.. Our lives and end it with “NO TAKE BACKS” and then we’re all fucked

      • @[email protected]
        link
        fedilink
        English
        33 months ago

        Write a program to infect the tastiest training data with malware that ends up turning off the system?

  • FaceDeer
    link
    fedilink
    233 months ago

    The list of features for the “ransomware” example he had ChatGPT help him write is almost identical to the list of features that would be present in a piece of software that was designed to help dissidents protect sensitive information from an oppressive government.

  • @[email protected]
    link
    fedilink
    English
    213 months ago

    This is like saying “you can write ransomeware in C++” and implying it is the language’s fault, somehow. ChatGPT is a tool, it does what the user asks it too (or it should, anyway). It has no agency nor morals. The argument that it makes it easier to write ransomware is silly as well. From high level languages to libraries and IDEs, we’ve always been developing tools to make programming easier! This is just the latest iteration.

    • Decoy321
      link
      fedilink
      English
      3
      edit-2
      3 months ago

      It writes code in the same sense that it writes sentences. Sure, it could look normal with functional syntax. But whether or not it functions is a different matter. People get these AI to write convincing nonsense all the time.

      It’s a great tool to help you get ideas, but it’s not a substitute for an actual thinking individual. Yet.

  • @[email protected]
    link
    fedilink
    English
    73 months ago

    I’d LOVE to see OpenAI’s lawyers’ reaction to that…

    Is liability a thing?

    "But, your honor, it wasn’t ME who wrote the ransomware, I only deployed it!

    It was ChatGPT who wrote it!"


    Yeesh.

    _ /\ _