• CosmoNova@lemmy.world
    link
    fedilink
    English
    arrow-up
    5
    ·
    edit-2
    5 days ago

    And even if it‘s true I‘m curious what it has to say about the Tiananmen Square massacre, Mao’s head count or the Corona pandemic. Huawei’s recent AI model that was developed for the western market failed to answer these and other uncomfortable questions with flying colors. Instead it gets very passive aggressive if you even briefly imply China isn’t the perfect utopia.

    • Packet
      link
      fedilink
      English
      arrow-up
      4
      ·
      5 days ago

      Doesn’t chat-gpt do literally the same in case you ask it on anything controversial? I think it is a quite basic thing for AI models.

      • CosmoNova@lemmy.world
        link
        fedilink
        English
        arrow-up
        3
        arrow-down
        1
        ·
        5 days ago

        It’s definitely not the same. You can be annoyed that ChatGPT sometimes says “It’s a complicated matter” then mentions some pros and cons but ultimately wants to move on to a different topic. It’s something else when the model says the pandemic originating in Wuhan is a huge lie and evil western propaganda. The Huawei model is very defensive and can quickly sound like an angry child. It’s painfully pro CCP.

        • Packet
          link
          fedilink
          English
          arrow-up
          2
          ·
          5 days ago

          Alright, gonna agree somewhat with ChatGPT, they do have one of the better lackluster censors. But even then, models like Bard exist. If you ask those anything related to the Israel’s genocide of the Palestinian people for example, it will simply refuse to answer. It is literally the same thing as with Huaweis model