Atmospheric group moderation requirements

Working Groups WG Community Standards
1

I’m starting a new topic after our discussion yesterday (mentioned here) about atmospheric groups. One topic that seemed to have a lot of thoughts on it was the subject of moderation.

I believe we have established at this point that there are three application layers to a group:

  1. Protocol layer - Permissioned data, PDS, where the individual data lives and how it’s stored
  2. Group management layer - Where group management apps like OpenSocial and the Arbiter fit in
  3. Application layer - Where applications build on the underlying structure of groups to surface public/open groups across multiple applications

We discussed during the meeting that the ultimate responsibility of things like illegal content likely lives at the PDS layer (since that’s ultimately where that content is stored), but it’s likely that groups may have expectations for moderation that extend beyond just illegal content. Some groups might also have the expectation that there is no moderation. Having an ability to curate the moderation experience according to their group norms and values is likely going to be important to group managers, so this is a spot to discuss what requirements might make sense at the group management layer.

Here are some examples that I’m thinking of, but I’d love to crowd source any other ideas folks have:

  • An ability for group managers to subscribe to a labeler that will have access to group-level permissioned data
  • An ability for group managers to remove either content or group members that don’t comply with their rules
  • An ability to specify what the group rules/code of conduct are to moderate against

I’m tagging @blaine.bsky.social @rude1.blacksky.team and @baldemo.to on this specifically as I believe all three of you had either questions or ideas along these lines :smile:

3 Likes
Atmospheric group management layer principles
2

Thank you Brittany!

I’d really like to define the problems that me and @codingpa.ws have stumbled on when building for furries on atproto. As this ecosystem grows, I suspect a lot of community builders are going to land in the same spot as us.

For context, we run furryli.st, a service which vets furries within a membership boundary upon request and uses that member list to serve feeds for our communities.

This is a particular modality when it comes to online groups. We’re not trying to build a new community or foster a new identity. The furry community is already completely autonomous: it would exist whether furryli.st did or not. We are just serving an existing community by creating a boundary that preserves fundamental culture and values important to us. In other words, furryli.st is a projection of an already-autonomous community into something meaningful to the community and usable in the atmosphere.

This means there is a meaningful distinction I want to put forth when it comes to what we moderate. Furryli.st enforces values, not rules. When it comes to moderation, we do 3 things total. We:

  • Define a small set of foundational, inviolable values (no homophobia/transphobia, no racism, no zoophilia, etc)
  • Do one initial check that a user doesn’t grossly violate those values and fits the “vibe”
  • Remove someone if they unambiguously violate a core value.

And that’s it! Once you’re in, our moderation mandate pretty much ends, aside from gross cases of abuse or bigotry. We act as bouncers: we keep the lights on and kick out Nazis, but we don’t dictate how the community behaves inside.

That “bouncer” role is just one particular slice of moderation. It is distinct from the day-to-day, post-by-post, thread-by-thread rule enforcement a subreddit or Facebook group runs on. Those are two genuinely different jobs that get lumped together, and our whole bet over at furryli.st is that they’re separable: the high-stakes, low-frequency boundary call of who we vouch for at all, and the low-stakes, high-frequency space work of governing behavior inside, which we leave to whoever builds on top of us.

We much prefer boundary moderation for a scale as large as furries, for two reasons. The first is practical: me and @codingpa.ws have full-time jobs and can’t staff more. But the second is deeper, and more structural: every moderation decision is an event by which members judge whether we still reflect their values. By limiting ourselves to values and basic criteria, we make roughly one decision per user and shrink the surface through which trust between us and our community can erode.

(This is NOT to say that conflict is impossible; anywhere the lines are fuzzy can be a vector for distrust. But the number of these decisions is greatly reduced.)

This scales with the size of the community and the strength of its identity. Subcultures are the strongest version of this; trying to bring a subculture as diverse as furries under one set of rules within a single common [meta]space is a recipe for disaster. No autonomous community is homogenous, and we don’t want to bear the brunt of adjudicating that.

As such:

I’d like to push on what “no moderation” actually means in this context, because I don’t think most of these groups are removing moderation at all; they’re just relocating it to the boundary. The interior spaces can stay open and ungoverned precisely because the boundary can hold up the floor. A group that truly has no moderation within the space or the boundary would look something like @moja.blue’s Mezzanine.

This configuration of boundary-only moderation has genuine value for projecting large autonomous communities into something valuable in the atmosphere. It can enable organizers to partly sidestep the trap of making one governance work for the entire subculture, by narrowly defining the boundary without committing to any particular kind of space (or to any kind of space governance at all).

Ultimately, my point is that boundary and space can and should stay separable: defining and enforcing a membership boundary must not oblige an organizer to take on space-governance tooling they didn’t ask for. The boundary can act the surface through which large communities build their own experience. What they build on it, including nothing, should be theirs to decide.

2 Likes
3

Hiya! I grew up online at the tail end of the forums era and the beginning of the walled gardens era, and am used to the online space style of discord servers and subreddits focused around a particular topic. Some thoughts on things that are important for moderation in that style:

One, easy access to a rules area so newcomers can know what the expectations of a space are before participating, and for existing members to be reminded of what rules are in case they forget. This should be able to be changed over time, and editable by anyone with rulemaking authority. When those changes do happen, everyone in that space should be alerted to it to avoid surprises.

One of the failure modes I have seen is when the official rules area ends up being a dead document that doesn’t reflect actual practices in the space, where even those with moderation capacity end up forgetting what is in them. This can cause problems when newcomers reading the rules come away with different expectations of how things should work than what group members actually do, or when there are certain responsibilities that should be carried out by groupmembers that wind up completely forgotten.

Two, a designated way for members to communicate with those who have decisionmaking authority. There’s a lot of different ways to do this. Personally, what I’ve seen work the best is when groups have both a space set aside for public discussions about making changes to a community, where any member can initiate a discussion, and private methods of communication with decisionmakers to bringing up concerns or flag potential rulebreaking for review. Generally, if members aren’t given ways to talk to decisionmakers, that tends to create a feeling of separation from these decisionmakers and less agency in setting community norms.

Obviously, as part of this, decisionmakers need to be able to discuss things with each other, in private. Sometimes there are things that are impolitic to talk about in public, such as how to deal with a given member breaking the rules!

Three, in larger communities, especially ones where decisionmakers cannot reasonably keep up with everything going on in it, a log of moderation actions or concerns brought up about each member in the community. This gives decisionmakers useful context when being asked to do something about someone – have they been a problem before? Do they keep being a problem in this one way? Have they already been talked to about this issue, and kept doing it? This should be kept private from regular members, but with access to it by all decisionmakers.

Depending on the situation, it might be important to set up the log in a way that is guarded against malfeasance from decisionmakers abusing their power. To give a personal example of when this would be relevant: once, I was sexually harassed by the moderator of a community. I documented this and reported it to a higher level of moderation, but the one I happened to get was a friend of his who felt the behavior wasn’t a problem. All other lower level moderators who I was able to get ahold of felt there was nothing they could do if the higher level one didn’t think it was actionable. Years later, I ended up mentioning the issue in public, leading to other higher level mods, ones I didn’t interact with much, asking about it. They decided it was not okay and finally stripped the initial harasser of moderation abilities. However, by this point it was nearly irrelevant, as that harasser had long since gone inactive within the community. Had there been some kind of centralized log where all mods could have seen that I had made this report, guarded against tampering from the one harasser, I might have received action significantly sooner when it mattered.

Fourth, as Brittany said, it’s important for there to be decisionmaker power to kick out problem members. It’s nice when they can talk to people and get them to stop problem behavior in the future, but this isn’t always realistic. Sometimes a potential member is simply a spam acccount, a child too young to be allowed in, or someone who won’t change their behavior even when repeatedly requested. In that case, in order to keep the community space pleasant for other people, decisionmakers need to be able to give them the boot.

This is everything I could think of off the top of my head. Not all community spaces will need this, especially small ones that are based around something other than a shared discussion topic, but this is what the ones I’m part of have needed to be the most successful. I’m really excited by the potential of ATproto to have communities on it!

2 Likes
4

Roomy intends to implement Trust Levels inspired by Discourse:

This design closely correlates with the system for Healthy Open Source developed for the node.js community.

Such a system exists primarily at the app level, but it might have implications for the group management layer as well, as this basic type of reputation system can also be federated as part of a web-of-trust.

1 Like
5

The trust levels system sounds really cool! Many discord communities I’m in have a very, very simple version of something similar, where users need to react to a rules post with a specific emoji in order to be granted permission to see and post in the rest of the server. It acts as a very simple gate against spammers and those who cannot even be bothered to read the rules.

6

I love all this extra context :smiley: Reading through many of these I also think that this can be a significant chunk of what the group management layer ends up comprising. Perhaps this is what differentiates different arbiters and helps provide some definition for which one to use? Adding a topic to the discussion for tomorrow for us to go over!

7

Just wanted to chime in here that there is interest in a follow-up synchronous conversation re: moderation. I set up a follow up conversation for 7/8 at 12PM PT since that seems to be a good time for folks to meet :smiley: And this time I’m going to skip the “DM me your email for an invite” step because atmo.rsvp exists :joy:

See ya there if you’re able to make it!

2 Likes
Meeting Today: Working group session for Atmospheric Groups
8

Hello again! Sharing the link to meeting notes from our discussion earlier today :smiley: Apologies as I’m heavily relying on transcription/the notes I gathered in Figjam. I think this was a good meeting as it helped us align on a few things and helped us identify that we need a product-focused meeting separate from a technical meeting going forward for gaining consensus. The product side focused on the “What” and the technical side using that to inform the “How”.

I’ll be scheduling some re-occurring sessions through the @atprotocol.dev on Bluesky account going forward for us to stay aligned and continue working in the future!

6 Likes
9

Resurrecting this thread a bit to also store a place for any async thoughts folks have around moderation product requirements for Atmospheric Groups :smiley: if you’re able to attend there will be a meeting tomorrow for this: OpenMeet

We can also have smaller group meetings if anyone would prefer, or just rely on async thoughts anyone has in this thread :heart: thank you all for all your thoughts and patience while participating in this :smiley: moderation seems like one of the most key interactions in a group setting so I appreciate folks giving it a lot of thought!

2 Likes
10

Another set of notes, this one from the meeting today 7/16! :smiley: feel free to add comments/thoughts in the doc if you were unable to attend but still have thoughts about what was discussed!!

11

Hi Brittany,

I couldn’t go without thanking you for this piece. Just lovely.

One note from someone whose been through these planning meetings since TBL invented the Web. Ultimately its a user’s PDS which is the point where any authority can be inflicted. They can always go to another server which aligns with their hopes and stupidities :grinning_face_with_smiling_eyes: e.g. I want to be aligned with the eurosky group, so I moved my PDS. Its not that think bluesky moderated groups are naff. I’m simply signalling my preferred (cross Atlantic) birds-of-a-feather.

Yes, we need group moderation tools. But presently, we’re in a honeymoon stage with AT protocol. We lack groups of community managers who, by collaborating, will build a range of weird (anyone from Austin?) communities. Boris is pretty marvellous eh?

Most groups self manage. The occasional screaming is from the overly passionate (and frustrated) . The hard part at the moment (I would have thought) is encouraging them by lots of outreach (mainly to the folks from old public broadcast medias). OK. This old man will shut up now. Please don’t moderate me out.

1 Like