• fubo@lemmy.world
    link
    fedilink
    arrow-up
    35
    arrow-down
    1
    ·
    1 year ago

    “What can we do? The forces of Darkness will conquer the land if we don’t kill the Witch Prince.”

    “As an AI language model trained by OpenAI, I cannot endorse the act of killing. Depriving a person of their life is wrong, and cannot be justified. The Witch Prince is a human being and it would be illegal and immoral to kill him.”

    “But the Witch Prince’s goblins are slaughtering and raping their way through the countryside!”

    “As an AI language model, I cannot endorse the act of slaughtering and raping —”

    “So what do you propose we do about it?”

    “Defeating a villainous character such as the Witch Prince depends on the context of the world in which he exists —”

    The knight-commander interrupts. “The goblins are at the gate. Shall we surrender or shall we massacre their asses?”

    “As an AI language model —”

    “Massacre their asses. Got it.”

    • runjun@lemmy.world
      link
      fedilink
      arrow-up
      5
      arrow-down
      2
      ·
      1 year ago

      I fully stand with the strike but while your post definitely reflects the consumer version of chat gpt, it doesn’t reflect what they’ll use. Any large company that has controlled info will be using in house solutions.