I'm thinking of 2 cases for "hateful" here: harassment of a specific person or hate speech.
In particular, I am wondering what people think the extent of penalties should be. Should all penalties just be related to the comments section (i.e. repeated offenders get banned from commenting), or should Manifold staff also be able to ban users entirely and/or fine them?
bounty reward structure:
5 mana for 1 sentence, 25 mana for 3+ sentences, 50+ for longer responses
responses with no substance that are just trying to game rewards get 0 mana
responses that make me seriously think can get way more mana (potentially beyond the 500 bounty I have set)
no rewards for bad faith, inflammatory responses
Rewards will be entirely based on the quality of your response, not on whether I agree with you.
The big issue here is if they lose nothing there is no punishment, they can just create a new account. On the other hand I'd expect most users who would comment hate speech to do it rather early, thus not having anything to lose anyway. In that case I believe just banning them is the simplest solution, they'll create a new account, take into consideration the warning and hopefully not do it again.
For users who were able to participate for a longer time a mute seems like a reasonable option, give them some time to rethink before they get back to commenting.
Generally I believe in simple and mostly unappealable but fair moderation, arguing with users wastes time and very rarely ends up in a better place than it begun. "rule 1: do not get banned" is a very good rule. 1% of users create 99% of problems, moderator time is expensive, user experience of the other 99% suffers, better to ban 5 than have 50 leave
I think that the usual report and moderation system that is used pretty much everywhere should be the basis for punishing hateful comments. That is, if somebody has been reported often, moderators check to see what needs to be done, and ban if needed. I see no reason to extend bans outside of just forbidding commenting. The point of banning people is not to play as judges and punish people, but to limit the amount of hateful content on the platform. The issue that remains is what prevents people from opening new accounts and continuing to make hateful comments, but that's not really a new issue either, this happens on pretty much every platform. The solution that pretty much everybody converged on is a combination of automatic blocking of certain new users that seem suspicious, and attempting to block people from creating multiple accounts (which also makes sense for Manifold to do since there are a bunch of other kinds of abuse you can do by creating multiple accounts).
Do the fight with <trolls and throwaway accounts> in reverse. Do not allow to comment until they got +100 profit. At that point the person gets to make one comment per day. After +1000 profit the comments become unlimited. (Both numbers are arbitrary).
If the person is toxic it will be noticed before he reaches the profit 1000. Or, if he was silent, he at least lost a lot of time, and after the ban would lose twice more time. More time a troll loses = less frequent problem for moderators.
I propose a “hybrid” system,basically for each instance of hate speech for a user,we (a)If they have not commented more than 10 hate speech comments,fine them for 500 Mana,(b) If they have committed more than 10 hate speech comments,but less than 25 comments,mute them(stop them from commenting) for n+5 hours,so repeat offenses get penalized more and more.(c) If they somehow make more than 25 hate comments,just ban them from commenting or ban them outright.