I agree that using AI to moderate without a human sanity checking it is a bad practice.
In many ways mastodon is superior for using humans, but honestly the bias afrom humans is a pretty big problem too. So im not sure mastodon is all that superior in the end, just a different set of problems arising from each approach and neither is great.
A better solution is to remove any sense of global moderation all together and create a system where the individyual users can be in control of what they see without having to put in the effort of moderating by hand case by case.
The way I'd achieve that is use a combination of AI and humans simply to flag accounts and servers as allowing certain content, have the human tagging be the entire population using a voting system. The tags dont do anything though, or moderate anyone. The users then can decide what tags they want to ban. This way the user is still in complete control of what they see and you dont have the issues with a purely human or purely AI moderator team.
GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.
All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.