I'm beginning to wonder if the only solution to hate speech and harassment on the Fediverse might be allowlist-only instances.
-
@tisha I think there are a lot of people who would be willing to accept that trade-off.
-
@RenewedRebecca The admins at @SocialCoop have experience running a social media co-op. I wonder if they have input on methods for assigning power to people, through votes or other means.
-
@jsit @timbray That would render it almost impossible to run a small or personal instance (you'd have to get onto the allow-list before you could communicate with anyone), forcing everyone onto fewer large instances, leaving us with "distributed twitter" rather than full federation. You'd likely get balkanization into a number of non-communicating clusters.
Also, both that "trusted group of people" and the managers of the resulting large instances would be exposed to huge legal liability. -
As a non-smart person, I don't have all the answers, but I am curious to know if this idea has been seriously explored before. It must have been, right?
-
> That would render it almost impossible to run a small or personal instance (you'd have to get onto the allow-list before you could communicate with anyone)
Only to communicate with the instances that set themselves up this way. And for many, this is an acceptable trade-off if it means stemming the tide of hate speech.
-
@parsingphase What would that legal liability be? I'm ignorant.
-
@jsit @ben @timbray
The idea has merit. spitballing - Maybe my instance can belong to multiple of these trusted rings of federation.Maybe getting in is like a handgun waiting period in the US, 5 day waiting period to federate, relatively small (25$ usd) one time deposit upon application unless you're sponsored by an instance/admin in good standing with the group. Both methods weed out the low effort harassment instances.
I could see our instance as part of a small self-hosted instance ring and also a regional instances ring.
A resulting question: who decides what behavior gets your instance kicked out of the ring, in a timeframe that allows a rapid response and defederation by all parties within the ring.
-
- Instance must be at least 6 months old
- Instance must have a code of conduct
- Instance's code of conduct be a superset of a core code of conduct
- $50/yr fee for inclusion
- One offense -- failure to ban an account known for abuse -- gets removal from list
- Can re-apply after one yearReminder that I have no idea what I'm talking about and these are probably laughably juvenile half-baked ideas
-
So my new instance is useless for the first 6 months?
Especially as a single user instance, or I need to get a self sustaining critical mass of users to join up front to be engaging enough for them to all hang around for the 6 month period. As they would be unable to follow anybody in that time?
Can't have an exception for single user instance, as once cleared I could open it up.
-
> So my new instance is useless for the first 6 months?
You would only be invisible to the instances that use this allowlist-only protection. This won't be every instance, of course.
Again, just spitballing here. Maybe it's a month, I dunno.
-
Raccoon at TechHub :Mastodon:replied to Jay on last edited by
@jsit @knizer @ben @timbray
The problem with things like this is new servers can't really show up and quickly proliferate under a whitelist system. That requirement of the instance being 6 months old would be a complete no-go, because it means you have to set up a server and go 6 months without access to whoever uses the whitelist.Meanwhile, whoever runs the white list is opening themselves up to all sorts of harassment that the average admin just isn't ready to handle, and that I would never recommend anyone take, because they would be left to defend the inclusion of everyone on this list. With a blacklist, it's as easy as pointing to a couple incidents with the server in question, but with a whitelist, the tables are turned, and any inclusion or denial is now subject to criticism.
This is why the people who run The Bad Space get so much shit whenever it blocks good faith instances.
-
I've been saying for literally YEARS now that the #Fediverse will evolve towards whitelists instead of blacklists. What you are describing is a whitelist and they have all the advantages you describe.
The downside is whitelists kill the dream of a single great unified Fedi. Instead you get N separate communities based on shared whitelists. (Possibly interconnected via hub-and-spoke instances.)
But there is no true way to fix the moderation issue in an open Fedi. Whitelists will happen.
-
Raccoon at TechHub :Mastodon:replied to Jay on last edited by [email protected]
@jsit @ben
The problem with things like this is new servers can't really show up and quickly proliferate under a whitelist system. That instance age would be a complete no-go, because it means you have to set up a server and then have to wait just to ask for access to whoever uses the whitelist, then go through the approval process.Meanwhile, whoever runs the white list is opening themselves up to all sorts of harassment that the average admin just isn't ready to handle, and that I would never recommend anyone take, because they would be left to defend the inclusion of everyone on this list. With a blacklist, it's as easy as pointing to a couple incidents with the server in question, but with a whitelist, the tables are turned, and any inclusion or denial is now subject to criticism.
This is why the people who run The Bad Space get so much shit whenever it blocks good faith instances.
(Continued)
-
Furthermore, these "trusted" instances could all relay with each other, vastly increasing the reach of posts on those instances within that group of instances.
-
Raccoon at TechHub :Mastodon:replied to Raccoon at TechHub :Mastodon: on last edited by [email protected]
@jsit @ben
I'm not saying this couldn't work for some servers, this might be a solution for curating the timeline better if everyone who isn't on it is effectively limited, or maybe for locking down servers during times of harassment or like right now when a bunch of neo-nazis have apparently made a bunch of new servers.It might also be useful for just small servers that are starting out and want to have a limited view of the network, though I would note that the problem of it not having the other small servers cuts them off from being able to chat with the people on those smaller servers, including the staff.
I'm not going to say that making a list of good actor instances is a bad idea, I think we could use it for something at the very least, I just don't think that most servers would really benefit from a bunch of people using a whitelist.
-
Jayreplied to Raccoon at TechHub :Mastodon: on last edited by
> That instance age would be a complete no-go, because it means you have to set up a server and then have to wait just to ask for access to whoever uses the whitelist, then go through the approval process.
Worth it to me if people stop getting assaulted with hate speech on Fedi. Not all or even most instances will opt to use this allowlist-only moderation, but people on those instances will feel a lot safer.
-
@Raccoon As for harassment toward the moderators of this whitelist and the instance admins who use it -- yeah, I have no solution for that.
-
Jayreplied to Raccoon at TechHub :Mastodon: on last edited by
> new servers can't really show up and quickly proliferate under a whitelist system
This is by design!
-
Raccoon at TechHub :Mastodon:replied to Jay on last edited by
@jsit
For reference, I've been harassed a couple times for suggesting servers to block, but I am harassed far more often for given reasons why servers should not be blocked. And I understand why people do that, because it's an assumption of bad faith when I give counterpoints or ask questions but that assumption of bad faith is not going to be good for whoever maintains this thing. -
Raccoon at TechHub :Mastodon:replied to Jay on last edited by
@jsit
Except new servers are the ones that are going to want to use a list like that, while older servers are more likely to have these built up block lists, as well as connections to servers that aren't on the Whitelist, and to have the experience and connections within the staff to be able to quickly identify and block instances.That may be a good thing, because it means that you can whitelist a bunch of servers that aren't going to ever use the white list, but once again, this runs into the problem of new servers not being able to connect with each other, and it would be a nightmare if any of those new servers became reliant on the Whitelist in the long-term.
All this being said, maybe you should actually put this list together: it might be easier to talk about how it might be used if we had a prototype to look at.