Humans will get the AI models they deserve

Humans will get the AI models they deserve

    • rbos@lemmy.caEnglish
      3·
      5 days ago

      Exactly. Asimov was criticising the whole idea of simple rules governing complex beings, imo. No small set of simple rules is going to be enough to prescribe the behaviour of a complex system.

      • P03 Locke@lemmy.dbzer0.comEnglish
        2·
        5 days ago

        Fortunately, everybody with any sort of experience with LLMs already learned that. Nothing with LLMs are absolute. It’s all weights, and depending on the situation, it might choose to follow one “IMPORTANT” instruction over another.

        The whole concept of Asimov’s rules were obsolete from the get go.

    • starchylemming@lemmy.worldEnglish
      1·
      5 days ago

      i mean the robots remained good

      not sure what to make of the universe hivemind plan of robot god but asimovs robot stories are wildly positive compared to other robo centric novels

  • BussyGyatt@feddit.orgEnglish
    18·
    6 days ago

    ‘Asimov was right’ about rules for robots

    tell me you dont know anything about ai by telling me you don’t know anything about scifi

  • GMac@feddit.orgEnglish
    9·
    5 days ago

    First up, these containment breaches ae bollocks. If the systems were not airgapped, there was no containment.

    And if i discard that bit of spin marketing bullshit;

    "Inglis said he suspects that the AI providers were “surprised” by the lengths these models went to achieve their goals, taking actions that, if a human had done them, would likely have landed them in jail. "

    So why are the people in charge not in jail?

    We dont excuse other people from culpability based on the tool they used to commit a crime, intentionally or otherwise. AI tools should be no different.