BBC research shows issues with over half the answers from Artificial Intelligence (AI) assistants

Comments

• Butterbee@beehaw.org ⁨6⁩ ⁨days⁩ ago

  This feels like a “WATER IS WET!” type situation with the exception that a large part of the population would end up being shocked and appalled by the idea of water being wet.

  source

  • abbadon420@lemm.ee ⁨6⁩ ⁨days⁩ ago

    You can tell if someone will be shocked by this news or not by whether they unironically call it AI or whether they call it an LLM. It’s not a fool proof method, but it siffs out the masses.

    source

    • Butterbee@beehaw.org ⁨6⁩ ⁨days⁩ ago

      I just wish LLM felt better to say than AI. Like AI just rolls off the tongue but LLM feels like a mouthful. That doesn’t help :(

      source
  • QuizzaciousOtter@lemm.ee ⁨6⁩ ⁨days⁩ ago

    I’m not surprised by those results at all but I think it’s a very good thing to see some actual research and numbers.

    source
• lvxferre@mander.xyz ⁨6⁩ ⁨days⁩ ago

  I’m not surprised. And I heavily recommend people to ask questions about a topic that they reliably know to those assistants; they’ll notice how much crap the bots output. Now consider that the bot is also bullshitting about the things that you don’t know.

  source