• @[email protected]
    link
    fedilink
    English
    113 hours ago

    Anti sematic behaviour is unfortunately part of the internet and general historical data, which means it could very well be an undercurrent in training data. AIs then could exhibit subconscious biases, no matter how many times the creators tell them not to be racist.

    • Singletona082
      link
      English
      210 hours ago

      Nono, the theory most people hear is the one about AI taking over. I read the original post with the theory written there devolving into this big giant global conspiricy where (coded language for jewish shadow government) is profiting off of everyone because of the AI takeover.

      The idea that human contnet is getting croweded out, that part i absolutely buy.