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.”

  • Metype @lemmy.world
    link
    fedilink
    English
    arrow-up
    26
    arrow-down
    1
    ·
    5 months ago

    So you have a product that you’ve made into a system for getting answers. And then you couldn’t be bothered to try and sanitize training data enough to get your answer system’s new headline feature from spreading blatantly incorrect information? If it doesn’t work, maybe don’t ship it.

    • xavier666@lemm.ee
      link
      fedilink
      English
      arrow-up
      5
      ·
      5 months ago

      I think the problem they are facing is data quantity. Sanitizing possibly terabytes of text data is a humongous task. They have probably used an AI to do the cleanup but the more suble errors have passed through the filter.

      • bignate31@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        ·
        5 months ago

        Yeah, the problem is how to sanitise effectively. You’ve gotta be able to find a way to automatically strip out “bad” things from your training data (via an “oracle”). But if you already had that oracle, you could just slap it on your final product (e.g. Search) and make all the “bad” things disappear before they hit the user (via some sort of filter).

        • xavier666@lemm.ee
          link
          fedilink
          English
          arrow-up
          2
          ·
          5 months ago

          I’m pretty sure google’s final solution will be using mechanical turks

    • iAvicenna@lemmy.world
      link
      fedilink
      English
      arrow-up
      5
      ·
      5 months ago

      The worst part is they don’t seem to realize their responsibility in this as the leading search engine that the majority of the world uses. They seem to have the mindset “our answers are potentially dangerous for users but it is ok we have an army of lawyers”

    • Rivalarrival@lemmy.today
      link
      fedilink
      English
      arrow-up
      2
      arrow-down
      1
      ·
      5 months ago

      Reddit users were infamous trolls and shitposters leaning heavily on sarcasm. The problem they are in facing is Poe’s law.