Instead of the current, AI-based system of content moderation, Twitter could experiment with different methods of community governance and judicial review.
Imagine a system where AI auto-censorship decisions could be appealed by staking some karma-points on the odds that a community moderator would support the appeal if they reviewed it. Others could then stake their own karma points for or against, depending on how they thought the community moderator would rule. An actual community moderator would only have to be brought in for the most contentious cases where the betting markets are between, say, 30% and 70% -- this would make the system more scalable since most appeals would get resolved by the community without ever escalating to a moderator.
You could then have multiple levels of appeals and judges, creating another market on whether some kind of Twitter Supreme Court would uphold the moderator’s decision. (The above idea is ripped directly from Robin Hanson but I can’t find the exact post where he describes it. It also resembles the dispute-resolution mechanism of the UMA crypto coin.)
Making nuanced, human-based judgement scalable in this way could both directly improve the quality of twitter discourse, and help familiarize people with an innovative new social technology. Also, by creating a system of community governance instead of AI-based censorship, it might offer a superior middle path compared to the current “AI-based censorship vs 4chan anarchy” debates about social media content moderation.
Instead of the current, AI-based system of content moderation, Twitter could experiment with different methods of community governance and judicial review.
Imagine a system where AI auto-censorship decisions could be appealed by staking some karma-points on the odds that a community moderator would support the appeal if they reviewed it. Others could then stake their own karma points for or against, depending on how they thought the community moderator would rule. An actual community moderator would only have to be brought in for the most contentious cases where the betting markets are between, say, 30% and 70% -- this would make the system more scalable since most appeals would get resolved by the community without ever escalating to a moderator.
You could then have multiple levels of appeals and judges, creating another market on whether some kind of Twitter Supreme Court would uphold the moderator’s decision. (The above idea is ripped directly from Robin Hanson but I can’t find the exact post where he describes it. It also resembles the dispute-resolution mechanism of the UMA crypto coin.)
Making nuanced, human-based judgement scalable in this way could both directly improve the quality of twitter discourse, and help familiarize people with an innovative new social technology. Also, by creating a system of community governance instead of AI-based censorship, it might offer a superior middle path compared to the current “AI-based censorship vs 4chan anarchy” debates about social media content moderation.