The "digital fiefdom" problem

[eblu@activitypub.space]

Recently, I've been keeping tabs on what people on Bluesky have been saying about Mastodon in order to understand the perspective of those that found the fediverse to not be a good fit for them. There are quite a few issues that I've spotted floating around (I definitely see an uncomfortable but necessary conversation about the culture surrounding the fediverse as a whole in the near future), but I really want to touch on the "digital fiefdom" problem because that seems to be the most prevalent one on both a social and technical level that hasn't really been touched upon yet.

Moderation and defederation can ultimately make or break someone's experience on the fediverse, especially when they want to follow a user on a different instance. It's certainly important to limit what comes into an instance for the safety of its users, especially as we have multiple confirmed cases of those that harbor people such as white supremacists and advocates for child abuse, however it's very easy for this kind of power to be abused and lead to very wide-reaching consequences for hundreds of users at the same time for reasons that could be as small as not liking the attitude of another administrator.

Let's run through what a typical scenario might look like for someone affected by this problem:

1. Alice signs up for an instance that they prefer.
2. Alice intends to follow Bob, entering their name into the search bar. Alice finds no results. This is obviously confusing.
3. Alice discovers that the instance that Bob is registered on is blocked on their instance at the server level, meaning that they cannot communicate with each other. Alice later finds that this was due to a disagreement with the administrators of both instances.

Now, there are a few courses of action that Alice can take here:

• Alice can attempt to set up a redirect of their account to a new one on Bob's instance, however they will lose access to their posting history as well as any relationships that they might have had on their previous instance. They may also lose even more relationships depending on what Bob's instance blocks.
• Alice can attempt to negotiate with the moderators on both platforms to lift the block, however this will likely not work and may even result in a ban if the moderator had enough of a trigger finger to warrant it.
• The easiest solution, and by far the most common: Alice decides that this isn't worth the headache and leaves the fediverse altogether, reconvening with Bob on a different platform and advising their peers against using it.

Needless to say, all of these approaches are less than ideal and I think it would be wise to see what we can do to tackle this issue sooner rather than later. Once again having the ability to block other instances is very important for the safety of others, and building "alternate pathways" would completely undermine these decisions, but maybe there is a solution where we give users more agency over who they are and aren't allowed to interact with?

Unfortunately I don't have a good solution for this beyond "just turn the fediverse into bluesky and make moderation entirely separate from instances altogether" but I do want to at least make the issue known so that we can start having a discussion about it. I look forward to seeing what all of you might have to say about this.

[julian@activitypub.space]

> typical scenario

That is not the typical scenario. Here's how it really goes (sorry to be blunt)

1. Alice signs up for an instance that they prefer.
2. Alice intends to follow Bob, entering their name into the search bar. Alice finds no results. This is obviously confusing.
3. Alice leaves the fediverse because she feels it is "broken".

Very few individuals would take the time to do the legwork to discover why they can't find somebody. The process is incredibly opaque, and I'm speaking from a POV where I actually understand the technical workings of AP.

---

Now, that part aside, the crux of the problem is:

> Alice discovers that the instance that Bob is registered on is blocked on their instance at the server level

In the grand scheme of moderation actions, instance-level defederation should be the last resort. There are multiple avenues of increasing effect that should be done instead. For example, user blocking, muting a user from another instance, putting that user's contributions behind a queue for manual approval, instance-level post queue, instance-level silence, instance-level block. Not all of these are likely available, hence...

The questions we should ask are:

1. Why were these actions not carried out?
2. If they were, why were they not sufficient?

Are the instances' Trust & Safety tooling insufficiently granular, and so the only option is to defederate?

[ulveon@derg.social]

@eblu@activitypub.space I wrote something similar about 9 months ago: https://ulveon.net/p/2025-01-21-fediverse-and-the-consequences-of-decentralised-moderation/

[julian@activitypub.space]

eblu I moved this topic over to a new Moderation & Server Administration category.

[eblu@activitypub.space]

julian This sounds a lot more in line with what's actually happening. Perhaps my perception has been a bit skewed doomscrolling through everyone who was unfortunate enough to land on/have friends on a badly moderated instance. 😅

As far as I'm aware I've constantly seen that people have been failed by Mastodon's lack of moderation tooling and it's definitely lacking granularity in actions that can be taken against instances: the only three options are blocking all media from there, hiding everything from public view unless you follow someone (Mastodon doesn't show you if a post was limited so this tends to break threads a lot), and outright rejecting everything from a server, so obviously this is far from granular and considering how overworked moderators tend to get they're way more likely immediately reach for the "defederate" option if they see something just a little bit off about another instance without investigating it.

I think it would be a good idea to have a bit more control on what limits you can impose on an instance that seems to be misbehaving? More collaborative cross-instance tooling would be nice, I have been thinking about "federating moderation activities" but I don't know enough about the protocol or web apps to really put anything down unfortunately. I think having Bluesky-esque external moderation services could really help out in the long run

[eblu@activitypub.space]

Just stumbled across a relevant thread on Mastodon's lack of moderation tooling, seems that there's a situation with a bad actor on the fediverse unfolding right now: https://disabled.social/@Aaidanbird/115286652333541455

Seems like there's also a case for how easy it is to ban evade and exploit poorly moderated instances but that's a bit outside the scope of this topic. It's definitely worth talking about though