Fediverse Moderation

From XPUB & Lens-Based wiki
Revision as of 14:47, 20 May 2025 by Wordfa (talk | contribs)

Here comes Moderation.

Examples

Manetta's suggestion :

https://github.com/f0x52/glitch-soc/blob/main/CODE_OF_CONDUCT.md

constantz :

https://pad.constantvzw.org/p/guidelines.questions

https://constantvzw.org/wefts/orientationspourcollaboration.en.html

port of Rotterdam code of conduct: https://www.portofrotterdam.com/sites/default/files/2024-06/code-of-conduct_0.pdf

old boys network 100 anti-theses

https://obn.org/obn/reading_room/manifestos/html/anti.html

varia code of conduct:

https://git.vvvvvvaria.org/varia/varia.website/src/commit/fe164708333a4707ac1103396ca331fc2cf2c509/content/pages/code-of-conduct.en.md

Research

Table 1: The prevalence of Rules in Mastodon and Reddit sub-communities. Rules can be both prescriptive and restrictive and can cover multiple content categories. For example, "Treat others with respect" is a prescriptive personality rule, "Don't be a jerk" is restrictive personality rule and "Be civil and be kind. If you need to criticise, punch up, never down. Hate speech, intended to incite or offend, is not acceptable" is a both prescriptive and restrictive rule that covers multiple categories, including personality, hate speech, and harassment. "Mastodon Instances" refers to those in the sample dataset (100 instances, 690 rules), and "Subreddits" refers to those in Fiesler and colleagues large-scale dataset [18].

from https://dl.acm.org/doi/fullHtml/10.1145/3584931.3606970

Rule Type Mastodon Instances Subreddits
Restrictive 93% 99.4%
Prescriptive 83% 99.8%
Hate Speech 71% 1.6%
Reddiquette/Mastodon Covenant 67% 5.9%
Harassment 64% 2.7%
NSFW 60% 4.1%
Personality 53% 30.4%
Content Warnings 50% N/A
Illegal Content 48% N/A
Consequences/Moderation 43% 90.5%
Spam 42% 1.7%
Incitement of Violence 34% N/A
Mis/Disinformation/Conspiracy 32% N/A
Doxxing/ Personal Info 31% 0.5%
Advertising & Commercialization 30% 7.6%
Images 28% 0.8%
Brigading/Dogpiling 27% 0.2%
Behavior/Content/Format Allowed 21% 99.9%
Impersonation 20% N/A
Codes of Conduct 17% N/A
Copyright/ Piracy 16% 1.0%
Server Interference 14% N/A
Automated Tools 14% N/A
Off-topic/topic Specific 12% 15.6%
Politics 9% 0.1%
Low-Quality Content 8% 0.8%
Reposting/Crossposting 8% 1.6%
Trolling 7% 0.3%
Links & Outside Content 7% 24.2%
Spoilers 4% 0.7%

Eldrich.cafe COC -> https://eldritch.cafe/about

  1. Hateful conduct: We do not permit any form of bigotry or hateful ideology, including but not limited to: fascism/nazism, racism, misogyny, homo/trans/queerphobia, antisemitism, ableism, as well as threats against marginalised communities or the “debate” of theirs right to a safe existence.
  2. Harassment: We do not permit harassment, including but not limited to: insults, bullying, dogpiling/brigading, bad faith trolling (“sealioning”), or boundary violations (for instance non-consensual flirting or unsolicited nudes). Requests to terminate conversations are to be respected.
  3. Safeguarding: We do not permit content that endangers others, such as others’ personal or identifiable information (“doxxing”) or the sharing of private conversations without consent. Promoting or encouraging self-harm or suicide will not be permitted.
  4. Sexual Content: Sexual content, such as erotic nudity or pornography, may be shared behind a content warning. Sexual content may not be shared on public timelines.
  5. CSAM/Abuse: Sexualised depictions of children, real (“child porn”, child sexual abuse material) or fictional (“lolicon”, “shotacon”), as well as sexual content shared without consent (reposting, “revenge porn”), are not permitted and will be removed.
  6. Misinformation: We do not permit the promotion of misinformation and conspiracy theories. Impersonating persons or organisations is not permitted, with the exception of parody.
  7. Spam: We do not permit spam, including but not limited to: excessive automated/identical posting, unsolicited off-topic replies, corporate advertisement, excessive/unsolicited promotion, crossposting on the public timelines. Bot accounts are not permitted unless administrators’ exceptional approval.
  8. Illegal content: This instance is hosted in France and subject to French law. As such, any content illegal in France may be removed without prior warning.
  9. Be nice: Mutual respect and kindness are strongly encouraged. Unprompted aggression may be dealt with as needed on a case-by-case basis. For the same reason, content warnings for common triggers are heavily encouraged, and may be required at the team’s discretion.

Drafts

https://pad.xpub.nl/p/Fediverse_Moderation