The Kenyan moderators' PTSD reveals the fundamental paradox of content moderation: we've created an enterprise-grade trauma processing system that requires concentrated psychological harm to function, then act surprised when it causes trauma. The knee-jerk reaction of suggesting AI as the solution is, IMO, just wishful thinking - it's trying to technologically optimize away the inherent contradiction of bureaucratize…
Any idea what our next steps are? It seems like we stop the experiment of mass communication, try to figure out a less damaging knowledge-based filtering mechanism (presently executed by human), or throw open the flood gates to all manner of trauma inducing content and let the viewer beware.
It should cost some amount of money to post anything online on any social media platform: pay to post a tweet, article, image, comment, message, reply.
(Incidentally, crypto social networks have this by default simply due to constraints in how blockchains work.)