• IndiBrony
      link
      English
      306 months ago

      Didn’t Microsoft do this and it just turned out expectedly racist?

      • @[email protected]
        link
        fedilink
        96 months ago

        Basically, yes.
        But with this question it might explore different … aspects of internet forums.

      • Kerb
        link
        fedilink
        86 months ago

        you mean tay?
        iirc it lasted 16 hours before microsoft had to shut it down

    • @[email protected]
      link
      fedilink
      66 months ago

      Just train it on the dark web. Lots of fun results with that, I’m sure lol. I’d love to see the chaos it’d produce. Within a controlled environment, of course.

        • I Cast Fist
          link
          fedilink
          26 months ago

          I’d like one AI manager to institute a weekday as LSDay, though it’s just as likely to institute a Kid Loving day 😫

      • I Cast Fist
        link
        fedilink
        56 months ago

        Wow… And if that alone wasn’t bad enough, https://www.inverse.com/input/tech/artificial-intelligence-4chan-bot

        Kilchner trained his chatbot, which he calls “GPT-4chan,” on approximately 134.5 million posts from 4chan’s /pol/ channel

        Trained exclusively on /pol/ posts is scraping the bottom of human stupidity

        Eventually, the bot had sown so much dissent in /pol/ that other users began accusing each other of being bots.

        Well, that’s a good ending