• dual_sport_dork 🐧🗡️@lemmy.world
    link
    fedilink
    arrow-up
    19
    arrow-down
    1
    ·
    1 year ago

    Alright, we’ll write a bot that can accurately moderate arbitrary internet content with an acceptably low rate of false negatives and false positives.

    You first.

    • cameron_vale@lemm.eeOP
      link
      fedilink
      arrow-up
      2
      arrow-down
      16
      ·
      1 year ago

      Here’s an idea

      When you read a post you vote it.

      This vote is also sticks to the person who wrote it.

      Whenever he posts, his post automatically get a (weighted) rating based on the history of your votes of his posts.

      Also, any post he votes automatically gets a (weighted) rating, for you, on his recommendation, based on his rating.

      This post voting rating propagates. And of course works for both positive and negative voting.

      Then you filter however.

      Everybody starts at 0. Which is also informative of course.

      • Ada@lemmy.blahaj.zone
        link
        fedilink
        English
        arrow-up
        9
        ·
        edit-2
        1 year ago

        That just means that folk from vulnerable minorities each individually have to downvote every new troll account targetting them, until the person just moves on to a new troll account.

        Which in turn is how you end up with communities full of nothing but white, straight middle class western cis men who think that trolling each other is a national sport.

        • cameron_vale@lemm.eeOP
          link
          fedilink
          arrow-up
          1
          arrow-down
          10
          ·
          1 year ago

          The cracking-resistance of this system is in the voters who are smart enough to vote as they like (flatworms can do it, so can we) and the depth and complexity of an organic voter/votee history, which would be hard to fake or quickly synthesize.

          Of course, yes, the proof requires pudding. A Lemmy fork? Ugh, it’s a lot of work. Maybe a friendly hs teacher can make it the class project.

          • Ada@lemmy.blahaj.zone
            link
            fedilink
            English
            arrow-up
            4
            ·
            edit-2
            1 year ago

            You miss the point. Your approach requires the targetted minority to experience the hate first, and then react to it, and gives them no method of pro-actively avoiding the content from new sources. It also ensures that every member of the minority in the community in question has a chance to see it, and has to individually remove it.

            That suits bigots fine, and unsurprisingly, isn’t sustainable for many targets of bigotry.

            • cameron_vale@lemm.eeOP
              link
              fedilink
              arrow-up
              1
              arrow-down
              7
              ·
              edit-2
              1 year ago

              Your approach requires the targetted minority to experience the hate first

              That isn’t so. There is vote propagation among peers to consider.

              If a trusted (upvoted) peer or peers downvotes a bigot (by downvoting the bigot’s posts) then you will see that bigot downvoted in your own perspective as well.

              • Ada@lemmy.blahaj.zone
                link
                fedilink
                English
                arrow-up
                2
                ·
                1 year ago

                You still see it though, especially if it’s a direct reply. And it is still a responsive system, that lets bigots just come back with new accounts and spew hate until they get downvoted in to silence, when they just come back with another account.

                Whilst the latter problem still exists even with moderators, at least a moderator can reduce the number of people exposed to hate.

                I’ve lived this. I have zero desire to use the system you describe, because I know it leads to toxicity that I don’t need.

                • cameron_vale@lemm.eeOP
                  link
                  fedilink
                  arrow-up
                  2
                  arrow-down
                  3
                  ·
                  edit-2
                  1 year ago

                  For older bigots you would filter them away.

                  For brand new bigots. That might require a “if the person’s history is too small, exclude” type rule. Which is less than ideal, yes. Lots of false positives there.

                  But let’s not put the cart before the horse. I think it’s a pretty good idea and I’d like to see it tested.

                  • Ada@lemmy.blahaj.zone
                    link
                    fedilink
                    English
                    arrow-up
                    3
                    ·
                    1 year ago

                    For brand new bigots. That might require a “if the person’s history is too small, exclude” type rule. Which is less than ideal, yes. Lots of false positives there.

                    Doesn’t work. For trans folk particularly, throw away accounts not linked to their main account is often the first step of exploring their identity online.