I mean that it’s gonna be inconvenient to consciously write down all the tags that apply, as opposed to the BCI giving a cloud of 2000 relevant tags/Discord reactions. It also feels like giving names would reduce the usefulness of this from telepathy to language.
The problem of specifying tags might not have received that much attention.
I had an expereince with ovewatch in that it is filled with all kinds of nasty behaviours. There were about 6 categories to report bad behaviour. Then you could also have free form text box to provide additonal details. Now free-form text you can’t automatically process that easily.
The reporting felt important so it pushed me to fight some of the inconvenience of speeding paperwork time on other’s behavoiural crimes. However the effort coming up stuff to fill on the textbox was kind of offputting. I ended up figuring out a way of “what is really objectionable about my experience?”. This made it so that in next bad experiences I could identify similar bad experiences with less thoughtwork.
Thus it became routine to write things like “Diagnosing others with mental issues (moron)” and “Diagnosing others with mental issues (autism)”. (And it leads to thinking whether “Diagnosing others with mental issues (idiot)” is a good thing to complain about (it isn’t it’s not diagnosing)). Abstractring them down to the objectionable part made me lose sight of the details which made them more directly comparable with each other. So the reports became really tag like and thinking about what the tags were made me think were the line of objectionability lies. If someone loses and is angry about it that is not objectionable. If someone bashes others and calls them bad names the motivation can be understandble but it is objectionable.
The system might have been designed more that it more relies on just adding up counts on the 6 fixed categories. And I seen a streamer file report where the textbox was just “fuck you” which speaks to players knowing the textboxes matter very little and that reporting is often done to express or play out anger. The lack of detail lost on the altar of automation has a significant cost. You could improve automation that it can handle more complex signals. Or you could evaluate that the social harm is great enough that more manual components are required. If such interactions were done face to face displaying any behavoiur that would be analogous would be very child like ot the range of 6 to 8 year olds and people would be directed to be more expressive and more constrcutive about conflict resolution. But online we are building places where that kind of culture is incentives or supported. Imagine what woud be the effects if in face to face communiation there was only 1 hostile communication available (no angry faces, or thumbs down or threatening voice or middle finger, just boiling it down to a a middle finger)
Create a tag dictionary.
I mean that it’s gonna be inconvenient to consciously write down all the tags that apply, as opposed to the BCI giving a cloud of 2000 relevant tags/Discord reactions. It also feels like giving names would reduce the usefulness of this from telepathy to language.
The problem of specifying tags might not have received that much attention.
I had an expereince with ovewatch in that it is filled with all kinds of nasty behaviours. There were about 6 categories to report bad behaviour. Then you could also have free form text box to provide additonal details. Now free-form text you can’t automatically process that easily.
The reporting felt important so it pushed me to fight some of the inconvenience of speeding paperwork time on other’s behavoiural crimes. However the effort coming up stuff to fill on the textbox was kind of offputting. I ended up figuring out a way of “what is really objectionable about my experience?”. This made it so that in next bad experiences I could identify similar bad experiences with less thoughtwork.
Thus it became routine to write things like “Diagnosing others with mental issues (moron)” and “Diagnosing others with mental issues (autism)”. (And it leads to thinking whether “Diagnosing others with mental issues (idiot)” is a good thing to complain about (it isn’t it’s not diagnosing)). Abstractring them down to the objectionable part made me lose sight of the details which made them more directly comparable with each other. So the reports became really tag like and thinking about what the tags were made me think were the line of objectionability lies. If someone loses and is angry about it that is not objectionable. If someone bashes others and calls them bad names the motivation can be understandble but it is objectionable.
The system might have been designed more that it more relies on just adding up counts on the 6 fixed categories. And I seen a streamer file report where the textbox was just “fuck you” which speaks to players knowing the textboxes matter very little and that reporting is often done to express or play out anger. The lack of detail lost on the altar of automation has a significant cost. You could improve automation that it can handle more complex signals. Or you could evaluate that the social harm is great enough that more manual components are required. If such interactions were done face to face displaying any behavoiur that would be analogous would be very child like ot the range of 6 to 8 year olds and people would be directed to be more expressive and more constrcutive about conflict resolution. But online we are building places where that kind of culture is incentives or supported. Imagine what woud be the effects if in face to face communiation there was only 1 hostile communication available (no angry faces, or thumbs down or threatening voice or middle finger, just boiling it down to a a middle finger)