FediForumTM

Virtual Unconference

@fediforum@mastodon.social

FediForum March 2024:

Fediverse Moderation: challenges, best practices, resources, and future directions

/2024-03/session/5-a/

Convener: Lillian Outlaw (@lillian@mastodon.iftas.org)

Notes-taker(s): Lillian, Jon (@jdp23@blahaj.zone)

Participants who chose to record their names here:

Tags / links to resources / technology discussed, related to this session:

Some of the key points:

  • need communication channels for moderators to discuss, get help, coordinate repsonses to bad actors and APTs – online forums, meetups,
  • multilingual / multicultural aspects of fediverse are wonderful but introduce challenges for moderators, how does a moderator who’s not German or an expert on antisemitism respond to a report of a German-language post as antisemitic? want to get help from peer group
  • shared interesting in promoting positive interactions
  • need better tooling, not realistic for mods to have to learn Ruby to scrub stuff from the database
  • best practices are emerging (e.g. shared vocabularies)
  • collaboratively sharing resources will be key

Detailed notes

What do people want to discuss? - We didn’t get to a

  • How to scale the bigger challenge of terrible people on the internet to a local instance – not just “bad stuff”, but also “illegal stuff”.
  • what are the opportunities using AI for moderation?
  • Ozone
  • (provisional) coalitions for info-sharing between server teams, especially w/r/t coordinated adversarial activity that can be difficult to detect without central telemetry

How do you promote positive interactions on the instance?

  • on a small instance: what shared instances do we use. common terminology allows for more relatable setup, strengthen the connection betweent he actions I take and what’s important to me in the real life. Run a Muslim-centric instances, so I’ll use Muslim-specific terminology in the server rules, makes it more intimate than just the broad set of rules.
  • multiple langauges and multiple contexts is a strength but also challenging. IFTAS is suggesting shared labels, there’s a lot of value, but also there’s concern about losing the instance-specific. Harassment for example, one instance says “we use this word for it”, but that’s harassment. When you want to take the contextualized rule and report it to somebody outside your space, what are the upsides of a shared common vocabualry so that we can get consensual reactions from a large community of volunteers – while maintaining the contextualization. https://github.com/iftas-org/resources/tree/main/LABELS
  • a tension as a moderator: being involved with and knowing your community, and being a person to the community, is important; but being a little separate from it is also important. Need to find the appropriate tension. Even in English langauge we have dialects and ability to code-switch.
  • running a server for a region which has a minority language (well majority in the region but minority in the country), focused on bilingualism, but doesn’t speak the minority language. social norm: people can write in either langauge and respond in either langauge … but it’s a challenge for moderators who aren’t bilingual. By definition, vast majority of mods are members of their community … they don’t want people outside the community to push buttons, but they do want input from experts outside the community.
  • do people think it’s going well on their instance? several yes’s, having a group of moderators (operated as a coop) makes it more acceptable, peopel feel there’s a check and balance
  • hoping that some way for mods to communicate, maybe even an instance for moderators. i get reports about users posting in german from a different instance, can’t assess.
  • part of why i started my own server is that i didn’t fit in well with others. it’s kind of just sitting here, there’s not a lot of growth, i’m spending out of pocket to keep it running … are people going to be using it? one of two known majority-Muslim servers, a different tack from is moderation going well, more “is the community going well” – pretty strongly no. info sharing: the movement of more organized adversarial activity into smaller zones online, the fediverse becomes vulnerable – spam, disinfo, etc. can’t image that there’s a single thing that addresses this, but what’s the constellation of orfanizations / groups / etc. What are people thinking about?

IFTAS is developing ISAC - focused on advanced persistent threatds, cybersecurity, etc. preparing to open a group that’s focused on that level, more on pipes and less about moderators. Migth be a group that could say here’s a range of IPs to shut down.

IFTAS Connect is a community space, multilingual, for each label a section, resources for trauma, also research including evidence-based tools and approaches

Could we leverage this as something more directly built into fediverse software. Like there are times that I get reports that are forwarded to me from instance admins that I don’t know why they were forwarded to me.

Tooling is very primitive, I’ve had to learn Ruby code in order to whack the database, can’t ask moderators to do that! With CSAM detection, once something’s identified everybody wants it gone.

German fediverse has a regular moderation meetup in the German Fediverse addressing such topics. Also, teaching some German people how to moderate as the Mastodon docs are in English.

Need to elevate the discourse: how to govern humans in the fediverse without an oppressive algorithm, interaction should drive the technology – technology should support, not impose. Example: Need to escalate this report to a peer who understands the langauge and culture. More generally: how do I talk across the ocean about this?

Collaborative resource sharing is going to be genuinely essential for fedi to thrive.

We’ve had decentralized spaes before, just not on the internet. How do you deal with threats like APTs – or just bad people, just organizing. Excited about organizations like IFTAS, known good actors, transparent … building trust is going to be key for solving these problems. Whether it’s doing data protection and privacy or dealing with coordinated spam waves. How much of this can we integrate into the tooling itself? One of the things that Bluesky’s doing well is providing a mechanism for sharing these tasks – of course like any of these things it can be weaponized.

@fihu: many months back, on the german fediverse faced several instances faced similar reports, so some guys from a big german instance called in a regular, informal meeting. We talked about bigger moderation issues happening around the German fediverse, and helped each other with moderation. Some folks did not know about specific moderation topics as the Mastodon docs are in English.

Random tangent: How can we better bridge communities from other platforms (e.g. there’s a couple large Muslims communities I’m part of on Telegram but it’s a struggle to get them to move to Yet Another App). large number of WhatsApp and Telegram groups, how to bridge in a way that helps with network effect problem but doesn’t ceded

Maybe we need to organise a “ModiForum” … could be something for IFTAS to do!

What about bridges? Is it best practices not to moderate technology? Do we leave it up to individual users? Mastodon isn’t the fediverse, ActivityPub isn’t the fediverse.

BACKGROUND

Short description of Today’s Open Forum:

Moderation on the Fedi isn’t easy, and there’s a lot of room for improvement. In today’s open forum format, we will ask moderators (and non-moderators alike) what approaches, tactics, and processes work well, what resources are useful – or, what resources would be useful if they existed. We will also discuss how Fediverse software could evolve to make moderation easier, more effective, and maybe even less stressful?

AGENDA:

  1. Introductions: Lillian Outlaw, Jon (Will include: work being done while Advising and Contracting with IFTAS)
  2. Overview of Fediverse Moderation (assuming everyone present is not a moderator!)
  3. Open Forum Questions (we may not have time to get to all of them, so if there is one you see here that you would like to discuss, or another not listed, please, let us know!

Introductions!

Lillian:

  • MSW from Fordham University, Licensed Clinical Social Worker; worked in mental health for over a decade
  • New to the tech-world, but believes that her experience will bring a unique perspective to the table
  • As a therapist with a background in Social Work, focus has been on fostering healthy communication, resolving conflicts and creating safe and supportive environments -– all of which is important to moderation
  • Eager to learn from the wealth of knowledge and expertise of those around me!
  • IFTAS: nonprofit providing support for volunteers on social web platforms
  • Link to IFTAS Website: https://about.iftas.org/
  • IFTAS Mastodon: @iftas@mastodon.iftas.org
  • Contracting and advising with IFTAS. Will be developing templates documents for a moderator code of conduct, moderator handbook, and member code of conduct.
  • We will also offer guidance and recorded training materials on how to use these templates effectively.
  • Jon and I have developed a survey to gather information for creating those documents for moderators!
  • Link to Moderator Survey: https://cryptpad.fr/form/#/2/form/view/0sU8JebkNU8lb1yTD0VFIFiX1HVXKB4r3At41z9gaNY/

Jon:

  • Strategist, entrepreneur, software engineer, activist, focusing on the intersection of justice, equity, and software engineering.
  • Runs the Nexus of Privacy newsletter: https://thenexusofprivacy.net/ and @nexusofprivacy@mastodon.social.
  • Privacy and civil rights activism and disinfo fighting
  • Interests include intersectionality, psytrance and darkwave music, and tarot
  • I’ve written about the Fediverse since 2010 and Diaspora. Mastodon, a partial history (DRAFT) and Don’t tell people “it’s easy”, and seven more things Kbin, Lemmy, and the fediverse can learn from Mastodon are two good examples.
  • Recently I’ve been writing a lot about Meta’s Threads and the Dediverse, including: Should the Fediverse welcome its new surveillance-capitalism overlords? Opinions differ! and Threat modeling Meta, the fediverse, and privacy

Overview of Moderation on the Fedi - we skipped this during the session

  • No central control: Unlike big social, the Fedi doesn’t have one company controlling everything. Instead, it’s made up of many smaller communities called “instances.”
  • Each instance sets its own rules: The people who run each instance decide what’s allowed and what’s not. This includes things like hate speech, bullying, or illegal things.
  • Instances can connect or disconnect: Instances can choose who they want to talk to. If one instance doesn’t like another’s rules, they can stop talking to them using a Deny List.
  • Users and admins handle moderation: People using the Fediverse and the admins of each instance are responsible for keeping things in check. If someone breaks the rules, they might get banned or their posts removed.
  • Some instances involve their users: On certain instances within the Fediverse, the people who use the platform are given a role in deciding what rules should be in place and what actions should be taken when those rules are broken. This could mean that users have a say in shaping the community guidelines, reporting content that violates those guidelines, or participating in discussions about moderation decisions. It’s a way to make the moderation process more transparent and inclusive, allowing the community to have a voice in how their platform is managed.

Open Forum Questions (We may not get to all of them!)

  1. Q: How do you promote positive interactions within your instance?

    A:

  2. Q: What approaches have you seen work well for moderation and community-building?

    A:

  3. Q: In your experience as a moderator, which resources have been the most helpful in supporting you?

    A:

  4. Q: Are there any moderation resources or tools that you enjoy using? If so, what makes them effective/enjoyable for you?

    A:

  5. Q: Think about a time when you successfully navigated a conflict within your instance. How did you manage that conflict and come to a resolution?

    A:

  6. Q: Reflecting on your experiences as a moderator, what experiences impacted you the most? Why?

    A:

  7. Q: What are the most valuable resources you rely on to keep your instance thriving?

    A:

  8. Q: Thinking about your positive attributes or strengths in your instance, what are they and how can you apply these qualities to promote a culture of mutual respect and understanding?

    A:

  9. Q: Looking ahead, what goals or outcomes would indicate progress towards a more supportive environment in the fediverse?

    A:

  10. Q: Can you tell me about a time on an Instance where you’ve seen resilience or adaptability (phrasing?) demonstrated.

    A:

  11. Q: Balancing moderation and inclusivity can be challenging. What insights or experiences have you had that helped you balance the two effectively?

    A:

  12. Q: If you were to design your ideal Fediverse platform from scratch, what features/functions would you prioritize to support moderators and users?

    A:

Fun Questions if time allows:

What’s the most amusing interaction youve had while moderating?

Whats the one thing you love most about being a moderator in the Fediverse?