As the US military begins integrating AI technology, simulated wargames show how chatbots behave unpredictably and risk nuclear escalation

Given that chat bot responses are a product of the training data that was used, it seems that Western media has an inclination towards fostering nuclear conflict.

  • nephs@lemmygrad.ml
    ·
    4 months ago

    But they're also trained in the English language. I wonder if a bot derived from other languages would behave the same.