• Optional@lemmy.world
    link
    fedilink
    English
    arrow-up
    32
    arrow-down
    1
    ·
    17 hours ago

    Turns out, spitting out words when you don’t know what anything means or what “means” means is bad, mmmmkay.

    It got journalists who were relevant experts in the subject of the article to rate the quality of answers from the AI assistants.

    It found 51% of all AI answers to questions about the news were judged to have significant issues of some form.

    Additionally, 19% of AI answers which cited BBC content introduced factual errors, such as incorrect factual statements, numbers and dates.

    Introduced factual errors

    Yeah that’s . . . that’s bad. As in, not good. As in - it will never be good. With a lot of work and grinding it might be “okay enough” for some tasks some day. That’ll be another 200 Billion please.

    • devfuuu@lemmy.world
      link
      fedilink
      English
      arrow-up
      5
      ·
      edit-2
      6 hours ago

      I’ll be here begging for a miserable 1 million to invest in some freaking trains and bicycle paths. Thanks.

    • Rivalarrival@lemmy.today
      link
      fedilink
      English
      arrow-up
      1
      arrow-down
      7
      ·
      7 hours ago

      It found 51% of all AI answers to questions about the news were judged to have significant issues of some form.

      How good are the human answers? I mean, I expect that an AI’s error rate is currently higher than an “expert” in their field.

      But I’d guess the AI is quite a bit better than, say, the average Republican.

      • Balder@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        ·
        edit-2
        17 minutes ago

        I guess you don’t get the issue. You give the AI some text to summarize the key points. The AI gives you wrong info in a percentage of those summaries.

        There’s no point in comparing this to a human, since this is usually something done for automation, that is, to work for a lot of people or a large quantity of articles. At best you can compare it to other automated summaries that existed before LLMs, which might not have all the info, but won’t make up random facts that aren’t in the article.

    • desktop_user@lemmy.blahaj.zone
      link
      fedilink
      English
      arrow-up
      1
      arrow-down
      7
      ·
      11 hours ago

      alternatively: 49% had no significant issues and 81% had no factual errors, it’s not perfect but it’s cheap quick and easy.

      • fine_sandy_bottom@discuss.tchncs.de
        link
        fedilink
        English
        arrow-up
        2
        ·
        5 hours ago

        If it doesn’t work then quick cheap and easy I’d pointless.

        I’ll make you dinner every night for free but one night a week it will make you ill. Maybe a little maybe a lot.

      • Nalivai@lemmy.world
        link
        fedilink
        English
        arrow-up
        4
        ·
        9 hours ago

        It’s easy, it’s quick, and it’s free: pouring river water in your socks.
        Fortunately, there are other possible criteria.

      • WagyuSneakers@lemmy.world
        link
        fedilink
        English
        arrow-up
        2
        arrow-down
        2
        ·
        47 minutes ago

        I work in tech and can confirm the the vast majority of engineers “dislike ai” and are disillusioned with AI tools. Even ones that work on AI/ML tools. It’s fewer and fewer people the higher up the pay scale you go.

        There isn’t a single complex coding problem an AI can solve. If you don’t understand something and it helps you write it I’ll close the MR and delete your code since it’s worthless. You have to understand what you write. I do not care if it works. You have to understand every line.

        “But I use it just fine and I’m an…”

        Then you’re not an engineer and you shouldn’t have a job. You lack the intelligence, dedication and knowledge needed to be one. You are detriment to your team and company.

      • fine_sandy_bottom@discuss.tchncs.de
        link
        fedilink
        English
        arrow-up
        7
        arrow-down
        1
        ·
        5 hours ago

        I don’t necessarily dislike “AI” but I reserve the right to be derisive about inappropriate use, which seems to be pretty much every use.

        Using AI to find pertoglyphs in Peru was cool. Reviewing medical scans is pretty great. Everything else is shit.

    • MDCCCLV@lemmy.ca
      link
      fedilink
      English
      arrow-up
      2
      arrow-down
      4
      ·
      11 hours ago

      Is it worse than the current system of editors making shitty click bait titles?