Elon Musk's Grok AI chatbot sparked controversy with antisemitic responses, revealing deeper systemic bias problems across major AI language models.
  • besselj@lemmy.ca
    0·
    1 year ago

    Is it really that hidden? If the training data predominantly comes from a website full of white supremacists, it will sometimes sound like a white supremacist.

    • spit_evil_olive_tips@beehaw.orgEnglish
      0·
      1 year ago

      the sources quoted/linked in the article:

      Maarten Sap told CNN.

      When CNN tested three popular AI chatbots

      CNN asked each chatbot

      KhudaBukhsh told CNN.

      Cameron Berg and Judd Rosenblatt from AE Studio wrote in the Wall Street Journal.

      OpenAI, which makes ChatGPT, told CNN

      it seems pretty clear that this “news” site is just summarizing articles from more reputable outlets so they can get their own ad revenue from them.

      and the cherry on top:

      After the controversy, Musk acknowledged the problem on his social media platform X.

      with a link to…https://x.com. not to any one particular tweet. just a link to the website.

      I had suspicions that this was AI-generated slop, but that proves it in my mind. no human journalist would embed a link like that.