You know how Google’s new feature called AI Overviews is prone to spitting out wildly incorrect answers to search queries? In one instance, AI Overviews told a user to use glue on pizza to make sure the cheese won’t slide off (pssst…please don’t do this.)

Well, according to an interview at The Vergewith Google CEO Sundar Pichai published earlier this week, just before criticism of the outputs really took off, these “hallucinations” are an “inherent feature” of  AI large language models (LLM), which is what drives AI Overviews, and this feature “is still an unsolved problem.”

  • Leate_Wonceslace@lemmy.dbzer0.com
    link
    fedilink
    English
    arrow-up
    5
    ·
    5 months ago

    only return results for a specific set of topics.

    This is true, but when we’re talking about something that limited you’ll probably get better results with less work by using human-curated answers rather than generating a reply with an LLM.

    • snooggums@midwest.social
      link
      fedilink
      English
      arrow-up
      5
      ·
      5 months ago

      Yes, that would be the better solution. Maybe the humans could write down their knowledge and put it into some kind of journal or something!

      • Excrubulent@slrpnk.net
        link
        fedilink
        English
        arrow-up
        3
        ·
        edit-2
        5 months ago

        You could call it Hyperpedia! A disruptive new innovation brought to us via AI that’s definitely not just three encyclopedias in a trenchcoat.