Ai Tries To Identify Toxic Gamers Via Language Doesn T Quite Nail It
An AI tried to figure out which gaming subreddits were the most toxic, and it doesn’t quite get it. Payday: The Heist? Really? Sure, it’s not exactly a family-friendly gaming community, but it’s also not exactly the most notorious one either. Let’s explain a bit: Clutch.win tried to use an IBM machine-learning AI to see if it could identify toxic gamer communities. To do that, it made the AI learn about gamer toxicity from Wikipedia comments, rating them on several categories including toxicity, obscenity, threatening speech, insulting speech, or comments targeted towards identity hate....