• CameronDev@programming.dev
    3·
    10 months ago

    I personally lean more towards humans for moderation, as words alone dont convey the full intent and meaning. And this cuts both ways, benign words can be used to harass.

    But of course, humans are expensive, and recordings of voice chat have privacy implications.

    • haui@lemmy.giftedmc.com
      2·
      10 months ago

      generally, yes. But computers can take care of stuff very well at this point. Kicking someone for using the N-Word does not need meaning. Just dont use it, even if it is for educational purposes (inside a game-chat for example).

      and recordings of voice chat have privacy implications.

      I dont think we live in the same reality. over 30% in the US use Voice assistants that constantly listen in to their conversatoins (was just the first number I could find, I’m not from the US). Having a bot in a game VC chat store 1 minute of text for 1 minute for reporting purposes is like 0.00001% of what is going wrong with security stuff. Billions of people are getting analyzed, manipulated and whatnot on a daily basis. A reporting tool is not even the same game, let alone in the same ballpark in terms of privacy implications.

      • CameronDev@programming.dev
        3·
        10 months ago

        Yeah, AI to knock out the egregious stuff (n-bombs etc) is prefectly reasonable. But there is still a lot of harassment that can happen the really needs a human to interpret. Its a balance.

        The privacy i am thinking of is the legal side of things. Google/FB/Apple are huge companies with the resources to work through the different legal requirements for every state and country. Google/FB/Apple can afford to just settle if anything goes wrong. A game studio cannot always do the same. As soon as you store a recording of a users voice, even temporarily, it opens up a lot of legal risks. Developers/publishers should still do it imo, but i dont think its something that can just be turned on without careful consideration.