209
Reddit is introducing a new moderator: AI (www.theverge.com)
submitted 1 day ago by to c/fuck_ai@lemmy.world
56 comments fedilink hide all child comments
you are viewing a single comment's thread
view the rest of the comments
[–] 13 points 1 day ago (2 children)

this has been happening for some time already. bans after saying a wrong word happen immediately within a few seconds

  • permalink
  • fedilink
  • source
  • hideshow 2 child comments
    •
  • [–] 8 points 1 day ago

    I got hit by a word filter too. Each individual word was SFW but the combination could be interpreted as hateful (even though it wasn't in context). So they have some sort of intent-matching system that is poorly implemented, probably piping it through the cheapest haiku model they can access.

  • permalink
  • fedilink
  • source
  • parent