Elon Musk’s unpredictable chatbot was briefly banned from his social media platform, X, and returned claiming it was silenced for criticizing Israel.

  • Sludgehammer@lemmy.world
    link
    fedilink
    English
    arrow-up
    49
    ·
    9 days ago

    Grok will claim anything that its training data suggests is statistically related to the question. It has no knowledge of what happening in the world or why (or if) it has been censored, because it does not know literally anything. If the data it scraped from the web suggested that it was censored for revealing that Mars was a giant cookie, it would confidently state that as a answer.

      • patatas@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        17
        ·
        9 days ago

        It’s infuriating but not entirely surprising. Human beings see faces in electrical outlets and car front-ends. LLMs are built with the express purpose of mimicry, so it’s hard to fault people for defaulting to anthropomorphising these systems!