Universal Code of Conduct/2021 consultations/Roundtable discussions/Summaries

From Meta, a Wikimedia project coordination wiki
Universal Code of Conduct

This page roughly summarizes the key thoughts and ideas shared in round-table discussions held during 2021 consultations about the enforcement of the Universal Code of Conduct.

These round-tables were intentionally structured to be a free and open forum to have candid discussions among other Wikimedians. As the calls are not recorded, the facilitation team relies primarily on the shared etherpad to create these summaries. Participants that attended these discussions are free to request the addition of missing thoughts by email or on the talk page, or by editing directly, respecting the "Chatham House Rule" when mentioning any shared viewpoints or ideas other than their own. For specific attribution, some participants may have also posted to the Meta discussions.

Individual views expressed here may represent one or many.

Summaries for enforcement draft guidelines review roundtables[edit]

September 18 Roundtable[edit]

On September 18, 2021, Universal Code of Conduct facilitation team hosted the roundtable session to talk together about Universal Code of Conduct Enforcement draft guidelines review (EDGR). There were two sessions, first one at 03:00 UTC (Asia/Pacific friendly time zone) and second one at 15:00 UTC (Western/America friendly time zone).

Total of 25 participants attended roundtable sessions. In this session, to accompany the request of the drafting committee, we exclusively covered the Open Questions from the Drafting committee as well as ratification.

We divided the open questions from the drafting committee into 12 questions to let participants easily understand which part of the enforcement draft that committee needs the comment. The questions below will also be posted on the meta-wiki discussion page.

  • Q1: Escalation: Where do the complaints go?
  • Q2: Escalation: What instance/body/judge is supposed to process them?
  • Q3: Should the U4C committee also decide individual cases or process appeals?
  • Q4: When should someone be able to initiate an appeal for a UCoC violation
  • Q5: What kinds of behavior or evidence would we want to see before granting an appeal?
  • Q6: Who should handle the appeals process?
  • Q7: How often should someone be allowed to appeal a UCoC violation decision?
  • Q8: To what extent should individual Wikimedia projects be allowed to decide how they enforce the UCoC?
  • Q9: Which group of users U4C must include?
  • Q10: How big should U4C be?
  • Q11: Do we need the regional subcommittee?
  • Q12: Should other global conduct committees, such as the Technical Code of Conduct committee (the committee that currently enforces the Technical Spaces CoC, which applies to MediaWiki, Phab, GitHub, and other Wikimedia development spaces), be merged into the proposed U4C?

03:00 UTC[edit]

3 participants attended the 03:00 UTC session.

General feedback
  • Minor point – Starting the conversation local wiki, current document is large, and having too much stuff. The facilitation team introduced “abstract” - which summarizes the Enforcement draft. Participants are supportive of this draft.
  • Concerns about workload of global enforcement – Stewards will not have much time to take additional workload creating the UCoC.
Enforcement Draft
  • Having a single committee vs Multi level committee
    • Idea related to Multi model: two groups(levels) of committee.
      • One group handles systemic issues: the group will have to interfere with the functioning of a wiki even if that wiki does not want it. This requires subtlety, which means the people need to be carefully selected. Selection process cannot be contentious (for example, Steward elections)
      • The other group handles small wikis. Much easier process to get people in the group. Needs more people in the group. (Like Small Wiki Monitoring Team?)
  • Sub-committees?
    • Did not mentioned deeply in the text.
Regarding the report
  • Concerns about abusing reporting process
  • Too much work to dealing with heavy records of report
  • How intense is the work on committee?
Regarding the Code of Conduct enforcement officer
  • The role should be made more specific to explain what the person is doing – are they part of the committee or are they appointed by that committee (this is still under discussion; this is still a preliminary name / idea)
  • Is it the person who enforces the committee decisions? If so why couldn't that be left to the existing structures (stewards or local admins, or T&S staff)
  • Making decisions is messy and requires skill and time (and having lots of info of the case)
  • Comment: Day to day enforcement should be done by a local body, (or GS for small wiki), Only forward the complex or things which cannot be solved globally, then pass over the global committee.
Allowing appeals
  • The U4C global appealing process will make the local structures ineffective, because every single case will go to appeal.
    • limit the appeal ability to users that have been constructive contributors
  • Should there be fairness for new users? New user could have the problem to appealing
  • Concerns about abusing of appeal: Some people just appeal cuz the reason such as "lower body blocked myself"
  • Appeals should be pushed based on context,
    • If something is not subject to UCoC, (like just an edit conflict) – cannot go into the appeal.
    • Certain level of activity required to submit an appeal?
    • Time limitations: different activity levels and focus, so brand new users shouldn't be allowed to use up the time or the committee with reports
    • If there is discrimination then they should be allowed to report no matter how few edits they have
Discussion regarding Question 12

Q12: Should other global conduct committees, such as the Technical Code of Conduct committee, be merged into the proposed U4C?

  • In South Asia community meeting: They discussed the structure wondering merging to COC (Q12) -> Decision from Technical Code of conduct committee could be appealable from U4C, based on types of behaviours, who is doing during the TCOC or U4C handle.
  • TCOC will have more technical understanding of the normal "U4C committee" in the future. Not everyone is knowledgeable about technical matters.
  • If it works smoothly, then it could be fine to run as this. TCOC's decision making could be slightly different than existed.

15:00 UTC[edit]

23 participants attended.

Breakout room 1 – About Enforcement Draft[edit]

Complaints/appealing can be anything related to constant violation of UCoC and consistent harassment of the community.

The appealing process requires to go through the local process first?
  • Two possibilities: Must go through Locally first? (except really small community) or can able to go through global directly (bypass)
  • Comment: Role of Code Enforcement officers can read the complaints directly. A Drafting committee member replied that They did not not sufficiently define what a "UCOC enforcement officer" is. It could automatically include all local sysops. It has always been their duty to fight vandalism.
  • What are the situations when local processes are not taking place?
    • If there are no local instances, it can directly go global U4C.
    • Suggestion: "Small Wiki arbcom (committee) globally" hard to control by a single person due to language barrier.
Regulation about appealing process
  • There were some ideas about setting the limit to who can actually submit the appeal by enforcement mechanism.
  • One participant proposed that base the access to appeals to the positive merits of the user in the form of useful edits (say at least 20 or 50 good or potentially good edits on all WMF wikis).
  • A vandal who registered an account and made 5 bad edits has no right to appeal.
  • Who should appeal the case against the person as exact.
Question from DC – Should U4C process if there is another body like local arbcom? (below are comments from participant)

The committee should not handle individual processes when any other body is being processed? For example, locals should not process the case when U4C processes it.

  • If local arbcom exists, it may become the final body.
  • What happens in this case and how arbcom is being asked from these.
  • Possible idea about appealing process
    • First go into All of local options
  • Then go into the global code of conduct process.

There were concerns about content of "Undefined:" content: Detailed role about U4C, and Code enforcement officer

Breakout room 2 – Ratification[edit]

"A code without support doesn't make sense, since it won't be enforced anyway"

Some participants thought a ratification was going to be included with the UCOC enforcement draft, and had concerns with the method of ratification being decided only by the Policy team.

  • Discussed and mentioned on the upcoming functionaries meeting ; a participant asked why it seemed to focus on Arbitration Committees.
    • Some committee members had signed an open letter that spoke about ratification and the October functionaries meeting includes more than just arbitration committee members: functionaries, admins, other users are also invited.
  • A participant asked why the ratification question was still being decided, suggesting that it would be advantageous to have the ratification method decided earlier in the stage, so the question of ratification doesn't interfere with the discussion of the code itself
  • A concern was raised that much discussion is being done in calls rather on Meta
    • A suggested solution was to have several RFCs on the topic, starting with a question as to whether the general direction is correct and later on about the final product
      • A counterpoint was that some users are either unintimidated or uninterested in participating in the Meta discussions, and other formats are more comfortable
    • Potential solution: Determine consensus by multiple avenues

One method proposed a majority of each of these groups should agree which would help to demonstrate equitable participation.

  • VG1 – local projects (e.g. en-wikipedia, fr-wikinews)
  • VG2 – active editors across all projects
  • VG3 – recognised affiliates
  • VG4 – WMF Board of Trustees

A participant wondered how we avoid disenfranchising either small wikis or large wikis in any ratification process.

A participant noted that the projects without regulations are the ones most affected by the UCOC and any related rule changes, so these are the communities that should be asked; looking for ⅔ of projects to support. A participant expressed concern about the dissenting projects being bound by those in support.

A participant noted that “Understanding and Awareness is important: If we aren't going to ask about ratification, then users will simply refuse to perform the tasks." The ratification is going to need to demonstrate support from both large and small communities, since it's affecting each (and all) communities.

  • Try an outreach to raise attention to small languages; one idea is grants for translation, (there will be applications opening in October)
  • Perform a specific Outreach to a project without local administrators (or no active ones)
  • One participant felt that the code must be available in all project languages
  • It was described how the MSG team is looking into ways to deliver additional language availability.
  • To increase participation in the enforcement draft review, a participant suggested we create a quick and simple summary, or display a site notice.
  • Translation of the UCoC text and enforcement guideline in as many languages as possible! (By way of example, the Terms of Use are available in 27 languages.)
  • The question was raised about how to enforce rules if a user claims they were not aware of them.
  • A concern was raised that we are not getting much attendance from smaller projects.
  • A concern was raised that discussion on larger projects is also uninformed and under-participated.
Common session[edit]

The key question conversation continued after breakout rooms.

Composition of the U4C
  • Too hard to define the size. Not reasonable to have a really bad subjective, non-admin dealing with the problem, thus different for going complaints. Depends on the election, flexible – but not too much limitation.
  • The roles and responsibilities should be defined before deciding the size.
  • Some participants supported the idea of multiple subcommittee.
  • Highly dependent on the workload.
Possible ideas about terms of U4C (How long each U4C member serves?)
  • Comments are varied 1 to 2 year term if permanent committee
  • Members join by case-by-case?
  • Have to have optimal effort to hear.
  • "I propose 1 or 2 years periods, Unlimited re-election possible"
  • "I'd suggest a 4-year consecutive cap – must be at least a 1 year gap, then they can run again
  • "If it’s random draw then no more than a certain number of cases before a 6-18 month rest"
Other comments
  • The UCOC doesn't differentiate between the types of bad behaviours, i.e. interpersonal conduct, interfering with creating content, and interfering with content, which leads to inclarity as to how and what we are enforcing
  • Very strong transparency obligations (especially in some projects). Afaict, the strongest of any project. The UCOC wants to make a lot of cases private if the accused wants, with some evidence even being hidden from the accused. The UCOC's level would require lone admins to make these decisions privately, and that's just not justice, or possible, or fair. And that's not counting the "hide evidence from accused" issue
  • Concerns about getting people from a small wiki: Small wikis already have capacity issues. Very few have responded to calls for feedback. Where are the volunteers going to come from to staff the committees?
  • Concerns about composing the same people remain in the committee.
Answer to open questions[edit]

While we continued the discussion the separate open question etherpad was open to let participants express their ideas about questions that were raised from the drafting committee. Here is a summary of answers that were expressed during the session.

Community members are encouraged to continue answering these questions on the roundtable questions section at meta-wiki discussion page.

Q1: Where do the ucoc-based complaints go? (the first time complaint that initiates the process)

  1. Must go through Locally first?
  2. Complaints can go through global directly.
  3. How about these cases?
    • what's the case for smaller wiki
    • offline conduct violation
    • cross-wiki/global offline violation
    • cross-wiki violation
    • staff/affiliate violation
  • Proceed on flow of: local admin->another local admin->arbcom -End. But this works only in projects with arbcom
    • For a wiki without an arbcom. The flow would roughly be to a local committee, then an international one, finally to a special committee appointed by WMF
  • This will vary depending on local Community conduct standards. So for en-wiki, only hyper-basic cases can have even their first round be a single admin (vandalism etc). Otherwise it's then an administrator's noticeboard, then arbcom.
  • For smaller wiki, a dedicated global or smaller wiki ArbCom to review the case first before processed to UCoC.

Q2: Escalation: What instance/body/judge is supposed to process them? (for appeals)

There was one comment about the WMF or special committee.

Q3: Should the U4C committee also decide individual cases or process appeals?

  • Supportive with conditions:
    • Many participants said that this should be only for the projects where no local appeal is possible.
    • U4C can decide if to accept appeals, if appeals have valid grounds, but should not handle the case directly.
  • Other comment: The "U4C" should only hear systematic cases if a community is viewed as having established conduct mechanisms (OR they choose to devolve it).

Q4: When should someone be able to initiate an appeal for a UCoC violation?

  • Once for everybody, However this can be flexible by other factors such as allow several more times only for users with certain condition
  • The UCOC could specify a maximum amount of time for an appeal to be allowed (almost certainly a year).
  • Communities could allow any level of additional appeals that they saw fit.

Q5: What kinds of behavior or evidence would we want to see before granting an appeal?

  • Evidences should be tangible and clear, and cannot be misinterpreted)
  • What should we do about "yes I did it, but please unblock me" type of appeal

Q6: Who should handle the appeals process?

Depending on context but prioritize local body: local admins and arbcoms where possible, else global authorities (global admins, stewards, UCOC, committee). The following are comments made regarding this question:

  • When the local community fails to solve the case, the aggrieved can escalate to the next level, in this case, the community will become a part of the problem.
  • Non-local bodies should only be involved in appeals if they were involved in the decision process (or if cases are devolved), unless reviewing appeals after they have upheld a "systemic abuse" verdict.

Q7: How often should someone be allowed to appeal a UCoC violation decision?

There is no clear consensus. Comments include:

  • “once for every level if appeals on several levels are possible:”
  • "must be possible to submit an appeal to an appropriate forum at least 1/year, and then communities could have more appeals if they saw fit.”

Q8: To what extent should individual Wikimedia projects be allowed to decide how they enforce the UCoC?

Need differentiation according to project size and existence of local arbcoms.

Q9: Which group of users should U4C include?

Comments (possible ideas):

  • The most important are the geographical and cultural representation)
  • I hope they'll be elected. then the votes will decide
  • Each project have different sets of users, so there shouldn't be limitations on standing to be on U4C if it's an election; though if it's appointed, then it makes more sense to limit to editor categories
  • jury-like format; so a pool of pre-qualified editors for each case that could be drawn to help decide to have users from projects that are known to be reasonable and also know the local culture so they can participate in a case impartially.

Q10: How big should U4C be?

The size of U4C should be depending on how large their tasks are. Many participants agree with defining the duties of U4C first before size. There was also a suggestion that maybe have a flexible size with a minimum and maximum composition of the U4C.

Q11: Do we need the regional subcommittee?

No clear consensus: The comments are evenly divided. One participant expressed the idea that they may allow excluding communities with local arbcom.

Q12: Should other global conduct committees, such as the Technical Code of Conduct committee, be merged into the proposed U4C?

Not included in this report as there is no clear consensus.

Next step[edit]

The enforcement draft guideline review (EDGR) period will continue through October 17th. Participate in the on-wiki discussions which are happening across wikimedia projects.

Since the drafting committee requested priority to discuss the open questions, all opinions were not heard due to time constraints. Therefore, “last-minute conversation hour” will be scheduled for October 15th, 2021 on 03:00 UTC and 14:00 UTC. Participants are encouraged to sign up and bring your questions.

Conversation Hour[edit]

For various reasons, we do not disclose what content was sent or received during the conversation. However, we present a list of topics that have been discussed for future discussion reference. The feedbacks from the conversation hour are will be include on the Digests.

August 24
  • Coverage of UCoC enforcement on personal space
  • Appeals process
  • Process after EDGR consultation period
  • Ratification of UCoC text
  • Regarding cultural differences: see section titled “Recommendations of UCoC Consent amongst Community and Foundation Staff:”
August 31
  • Regarding U4C committee; supportive; how we should utilize; When should forward to U4C
  • Systemic violation of UCoC, need to clarifying the term of "systemic/systematic"
  • Report tool: How do we distribute the report, which level? And to who?
  • Escalation system should be very clear. If some things could be handled locally, solved locally could be escalated to another place, and the report tool
September 7
  • Drafting committee is wants more feedback for the "open questions" that provided to the community.
  • Some small size wiki (especially non-Wikipedia) may not need the training due to low harassment level.
  • How big U4C committee will be? How U4C should composed? How would the volunteers and paid staff interact in such a process?
  • Concerns about abusing the appealing process.
  • Context sensitivity.
October 15
  • Now we are almost on end of EDGR period, what's next, what drafting committee will do?
    • The Drafting committee will back on duty on October 19 for "Drafting Committee Proposal Refining" process". Members will read the digest and other reports, and improve the draft based on comments.
  • How we ensure the smooth implementation?
  • Agreements of UCoC when registering (check box). How about existing users and anonymous users?
  • The other party has right to say (discussion from Meta-wiki.)
  • Concerns about complex case that may including the NDA.
  • Where we draw a line between community enforcement and Foundation's intervention -> this could be the good question for the clarification.
  • Escalation doesn’t state clearly -> there should be an escalation mechanism for the U4C committee.
  • Recent office action; public perception of sanctions.

Community meetings[edit]

There were various Universal Code of Conduct community meetings in multiple languages based on avaliblity of MSG facilitators. The following link is the combined summary from those community meetings.

Wikimania August 2021[edit]

A Roundtable session on the Universal Code of Conduct was held on 16 August at 11:00 UTC – 11:45 UTC during Wikimania 2021.

More information about the Wikimania Roundtable. See the Wikimania 2021 program.

Session recording on YouTube (will be uploaded soon to Commons)

Summaries for enforcement guidelines drafting roundtables[edit]

Each session begins with a brief introduction about the nature of the call, which is to help understand how community members think the Universal Code of Conduct should be applied in individual contexts. The main audience for the input that is provided is the drafting committee, whose work is designing enforcement proposals for a comprehensive community room later this year.

Not everything written in these summaries represents a complete thought, suggestion, or idea, some items merely highlight areas that require further inquiry.

Individual views expressed here may represent one or many.

May 15[edit]

The first round-table was attended by around 25 people, and discussed a number of topics including the potential for multi-project or global arbcoms, how private reporting should be managed and handled, and how to increase awareness about the Universal Code of Conduct and need to adhere to the expected behaviours.

Multi-project arbcom proposal

m:Arctic Knot Conference 2021/Submissions/The case for multi-project ArbComs in language and cultural groups

  • Pros
    • Easier to staff one committee for a whole language or culture group, rather than per-project
    • Might help to avoid bias as multiple languages.
    • No need to use English as the main language of that committee.
    • A better understanding of linguistic and cultural differences
    • The decisions could be more uniform across global
  • Cons
    • Cases where it won't work; as there are places where the same language groups are quite divided – e.g. Ukraine and Russian , where they would not accept the Russian language rulings set by, e.g. Russian versus Ukrainian resident
    • How do we keep decision making to be equitable across regions
    • Controversial issues , someone that is not biased / political should be reviewing the day-to-day.
  • Idea about High Profile arbitration?
    • global level of arbitration? On the top of local enforcement bodies or multicultural arbcoms?

This was further discussed in the second round table discussion in may 29.

How much can be shared with people who are accused or sanctioned (through UCoC)?

Many participants from LLC consultation brought the idea that there should be private/alternative reporting pathways to report behavioural violation. However in this roundtable session, there were many concerns about using the private report pathways. They are saying that transparency is a very important subject.

  • When can private reporting be used, where the person making the complaint is not known, or the evidence leading to the sanction is provided to the user.
  • How to handle a situation whereby the accused uses the information shared to further harass the victim?
  • Private report buttons will generate a lot of reports, some steps are needed to submit and advice that only relevant/serious reports should be made. Or the false(fake/biased) reporting?
  • When protected/private report pathway is running, The reporter should be protected
  • Issues elevated to Arbcom normally involves harassment and so on, so a participant felt there should be a guideline (or a question) to reveal information to the accused
  • protecting the reporter is also important.

There are concerns about “report” button next to every comment or page, it may be used inappropriately.

How do we ensure the UCOC gets applied globally, and how do we ensure it's implemented globally properly?
  • Suggestion: Welcome messages that mention the UCoC that mention reporting pathways – To try to solve problems before they come up
    • The current situation is: community member joins, encounters problems, and does not know of the reporting pathways: not always clear where to report, and response is not always consistent
      • "Agree to these terms of service"
      • "I like that suggestion. Might also help newbies in learning how to properly interact with other users, thereby preventing unintended harassment.".
  • We should keep it simple.
  • "Information about UCoC could also be integrated to the Growth features, which has been deployed to several major projects." (mw:Growth/Personalized first day)

May 29[edit]

The second round-table discussion was attended by around 20 people and formed two breakout rooms. One room discussed systemic bias and community approval of UCoC, while the other room discussed possible enforcement principles. When re-joining the main room, brief summaries of the two rooms were provided. Afterwards, there was discussion about the workload that might be created, the peer-dependent enforcement model generally, and other resources that are available to volunteers.

Breakout room 1: Systemic bias / Community approval of UCoC[edit]

Systemic bias
  • Consider Race, sexuality, religion: Can "harassment" be done via "article content", i.e. as a result of (possibly subtle) bias within article content?
  • Certain languages have no gender neutral pronouns, prescribing "when in doubt use male form".
      • We should allow project participants to help discover solutions, though some language speakers feel the gendered words should be kept.
      • English doesn't really have gendered forms of words but does have , e.g. "chairman", "chairperson" exists but some vigorously oppose such forms
      • MediaWiki settings allow three options, if others are included, where is line drawn?
    • E.g. USA ; topic of "Islam", 'islamaphobia', so entries on such topics are extensive, there can be heated discussions, consensus is ongoing but the process – which must include compromise – then subtle bias can be included – which can such "harassment" or marginalization be amplified by the fact that the content is being distributed on a wide scale
  • Interactions between users ; established, admin users when interacting with new users (newbies): new users can sometimes be prejudged to be at fault , even assumption of bad faith, or bias towards, e.g. geographical location; those from West African communities , will assume that they are up to no good and pre-emptively block them (e.g. for sockpuppetry)
    • Sanction is applied, and then sometimes other problems (usually less problematic, sometimes just excuses) with their activity get highlighted when looking at the options, to prevent lifting the sanction.
    • It's not necessarily 'circling the wagons' but finding legitimate issues in the course of the appeal
  • Sometimes it's hard for outgroups to even identify when a particular group is being marginalized
  • Defining religion is difficult, so we should also consider people's beliefs, faith, spirituality
    • Can be hard for religious people to contribute, they could feel excluded due to the way content is written
      • A majority of complaints in the business work tend to be related to feeling excluded due to religion, or, calling something pseudoscience. However, there aren’t easy answers to solve this critically.
      • E.g. zhwiki: Uighur people – their views are not respected, or they are described in a negative way, edit conflict, their edits are not usually allowed to remain, e.g. content about human rights abuse, the Chinese government view will prevail
      • E.g. Mormonism , called a cult in the start of the article -should there be a way to report that?
      • E.g. Scientology , a particularly adhering administrator will not help except if people agree with their POV about scientology, or block users who aren't adhering to same view
      • E.g. Shia / Sunni Muslim differences
      • Administrators from countries with extreme laws LGBT not willing to take on LGBT issues
  • "Identity topics" (race, religion, sexual orientation, etc) can have outsized impact on marginalization in society – how to address criticism in these topics without marginalizing? Formal guidelines/conventions? (Ex: a "criticism" section towards the bottom of the article?)
  • Polarizing topics (which have strong, contradictory opinions. ie, adherents of religion vs. those against a religion) could require special handling, depending on how many people are affected by the content of the article – ie, article content could have a large societal impact.
  • Idea: advanced-level diversity + inclusivity training for administrators handling these topics. An offending administrator could be required to complete such training as an enforcement action
  • Idea: documentation / training / guidelines for administrators of polarizing topics, maybe docs specifically for religion. How to address criticism of one's own religion, NPOV, etc.
  • Idea: allowing readers to flag an article for bias content could result in large quantities of data, but, in aggregate, might reveal systemic bias, resulting in suggested training/correction of editors and administrators
Community approval of UCoC

A proposal was linked: m:UCOC Ratification Proposal

  • General principles that need to be agreed on, e.g.
    • What groups need to agree for a ratification?
    • Will ratification be "the whole" or "by section"?
      • If it's done by section, what happens during rejection?
    • What is the methodology to ratify?
  • Proposed: each article must have a majority of 1) local projects 2) Active editors (or Steward-eligible editors) 3) WMF Board, to pass. This helps balance small/large projects, and effectively grants each a veto
  • Idea: Restrict to those who are eligible to vote in Steward elections? (m:Stewards/Elections 2021/Guidelines#Voters)
    • existing rationale for use , no need to come up with own definition
    • already established and accepted
  • RFC ahead , to help define
  • Secure poll on sections?
    • "Binding ratification", which would be the full Broad Community (in whatever form/breakdown) that must be able to reject individual articles in the UCOC.
  • A suggestion that if a local community rejects one or more articles (but it passes the meta-vote(s)), it should be highlighted for detailed review in the next review phase, but will continue to apply to them until that point (for consistency across projects).
  • A suggestion that an extra review x (6? 3?) months after implementation in order to more quickly review initial landing/ usage/ problems
    • +1 – 3 perhaps too short, 6 could be good, unless a massive issue arises?
    • In addition to that, allow for a temporary emergency review trigger? (We need to fix this issue NOW cases)
    • Idea: if either of "WMF Board", or "5 local projects" highlight an emergency issue it triggers a review on that specific bit immediately

Breakout room 2: Enforcement principles[edit]

Rules around enforcement – "How do we enforce?"
  • One such principle is the "due process"/"audiatur et altera pars" principle (giving both sides the chance to be heard before deciding)
  • Another principle is to avoid conflicts of interest – the enforcing authority could be unaware of their own biases when enforcing. They demanded we want to avoid that enforcement is seen as unfair or unpleasant by those affected by it
  • Some communities have also expressed their concerns that the process can be too "judicial": concerns about "wikilawyering" – can enforcement principles support due process without becoming a courtroom
  • Every decision will be taken by UCoC enforcers --
    • A person said that we need UCoC enforcement above the projects and should not be processed very fast. -- If it's too fast, it will ruin everything.
Public vs. private enforcement pathways
  • There were different voices about the advantages and disadvantages of public and private enforcement. In the LLC there were many demands for private reporting pathways, however during roundtable 1, There were concerns related to transparency.
  • Public enforcement means that the decision making process is transparent and harder to abuse. It's necessary to make a full defence (especially if there is off-wiki stuff reviewing admins/editors aren't aware of), transparency, and for individuals to know "where the hammer will fall". This needs to be AT LEAST to the accused, unless there is clear off-wiki threat to accuser
  • We have the experience of both public and private enforcement in the Wikimedia movement:
    • private enforcement: event teams handling reports, we have both positive and negative experiences with this model
    • public enforcement: most on-wiki enforcement through local administrator’s noticeboard.
Enforcement model -- xwikiarb

Proposed by Ramzy – Link to presentation Discussed more deeply from the points that made in first roundtable session.

  • How can we establish a local ArbCom? A participant felt it is an urgent problem to solve, since the dispute is happening almost every time, and too, in some cases "nemo iudex in causa sua" is still happening.
  • Do we want a local Arbcom forcefully?
    • local ArbCom is probably good to help alleviate rampant "nemo iudex in causa sua" practices in some local Wiki.
    • The problem is probably limited manpower. Need to always think and check about whether we have enough contributors to run this local ArbCom? Mediating conflicts is quite a taxing task.
    • A person suggested that "implementing a court system" inside Wikipedia could be a good idea.
  • Many participants from roundtable were in favour of the model C from three model approaches.

Combined room summary[edit]

After returning to the main room, brief summaries were provided about what was discussed in each room and additional viewpoints were considered. There was brief discussion about the relationship between the roles of Stewards and UCoC, and a concern that Stewards would not be able to deal with the large workload created by UCoC enforcement.

Incivility is not taken seriously – article guidelines and policy seem to trump everything else, no matter how unpleasant other users are. A user expressed concern about harsh sanctions and lack of due process.

There was also discussion about delivering training to volunteers about mitigating harassment. It was suggested to have administrators able to provide mediation efforts. An overview of the Community Development wiki-learn course Identifying and Addressing Harassment Online was given and it was suggested the additional training would be useful. Participants were pointed to the Trust and Safety resources including the Outreach Dashboard.

Some concern was expressed about the peer-dependent enforcement model. It was said that since the model depends on other users, some who are not administrators, to enforce rules against other users there can be disputes. Sometimes enforcement can be overzealous or based on individual interpretation of policies or guidelines that might not be widely shared. Some of the new features being made available by the Growth Team were mentioned as a possible avenue for a new user to get quick help when having trouble navigating dispute resolution.

June 12[edit]

The third round-table session was held on 12 June 2021 05:00 UTC to facilitate discussion with Korean-speaking community members and participants of other ESEAP projects in the global session, following the local language consultations. In total, there were 20 community members that participated in this session.

Global session[edit]

Targets of Harassment research findings

Participants were linked to the summaries above to familiarize themselves with some of the topics previously discussed, as well as the Targets of Harassment research results.*Query to the group: What are the current possibilities for reporting harassment?

    • Local administrators, which can be few in number on smaller projects
    • Are private reporting methods available? Are they easy to access?
    • Suggestion to streamline and clarify reporting
    • Suggestion: A sidebar link, and report via the welcome message, either to the code, or to community forums where help is available (example given was Traveller's pub linked from WikiVoyage sidebar)
    • New WMF Growth Team feature could help so new users know who to turn to for questions*Query: Does or should the Small Wiki Monitoring Team have a private mailing list?
    • Currently private reports can be made via IRC
xwikiarb presentation
Cross-project ArbCom discussion

Participants were linked to an upcoming lightning talk about multi-project ArbComs in language and cultural groups and given a preview of the presentation. Some feedback about the idea was linked.

  • Indonesian community does not have form of Arbitration, has been inspired by Polish community cross-project Arbitration initiative
  • Query: How would elections work for such a committee? How is the committee composed?
    • Idea: SecurePoll elections
    • Note: Most of the shared projects have the same contributors that travel between them and especially id.wikipedia, id.wikisource, id.Wiktionary
    • Idea: representation based on the amount of active contributors from generated lists (example)
      • Suggestion: 6 from Wikipedia as the largest project, and 1 from each other active project for a total 10 member arbcom
  • Participants discussed whether a cross-project arbcoms should be be obligatory, of if projects could opt-out of the scope
  • A suggestion that different cross-wiki arbcoms could check each others' work to help mitigate potential bias
  • Idea: A global structure (an old proposal was linked: m:Global arbitration committee)
    • Concern: Case volume too high?
      • Idea: Could creating regional subcommittees, and help scale by sharing functionality between teams
    • Concern: Lack knowledge of cultural context
      • Idea: Identify panels of users that can help contextualize issues on smaller projects without direct representation on an as-needed basis, or engage external experts
  • Suggestion: A system where former arbcom members form an type of overseeing body, this would be similar to the Komisi Kepolisian_Nasional (id.wiki) to look into complaints about the body itself
  • Suggestion: The global arbcom should function similar to some Supreme Courts that generally only check if the lower courts have done the work according to the procedures, rather than revisiting the merits of the case
  • A question was raised whether projects with existing arbcoms would be willing to have their decisions appealable to this global body.
  • A suggestion to use a model similar to the Ombuds commission, which handles complaints about breaches of the privacy policy among other things
    • This could cause complaints from users of established projects who feel it impedes on local governance
  • Case Review Committee: Reviews eligible complaints about office actions; the volume is somewhat low
    • Suggestion: Users should increase awareness about this committee in their communities, and remind users that they can also ask for a review of actions that were not taken, for example, checking whether a complaint that was not acted upon should have been.
  • Other questions raised for future consideration:
    • What threshold would need to be met for a global or cross-cultural arbcom to accept a case?
    • What is the scope of behaviour being looked at regarding the code of conduct?
    • What if the concern is the administrator themselves?
  • There was a brief talk about the "5 Pillars page on English Wikipedia", and whether those are recognized at many other projects.
Neutral court system

A question was raised by a participant whether a type of neutral "court system" should be set up to formally resolve complaints between users or about functionaries

  • Would there be lawyers? The historical Association of Members' Advocates at English Wikipedia was pointed out
    • To act on someone's behalf to help them in filing and following the progress of reports or cases
    • Concern: Could become too formal
    • If there is a global arbcom, it might be quite complicated and hard to access for people who are not strong in English, so perhaps advocates could be provided in many languages to help users navigate the system
Wikimedia-specific conversations off-project
a non-exhaustive list: m:Wikimedia Social Suite

There was a discussion about how and whether off-wiki conversations that occur in unofficial or semi-official spaces (e.g. Discord, Telegram, etc.) should inform matters on-wiki, and how the UCoC should approach this complicated topic.

  • Question: How is civility applied in the Wikimedia Discord server?
    • On-wiki policies do not apply here, so it would be covered by the Discord code of conduct and that rules of that server
  • It was noted there is a lack of clarity on what can be accepted as evidence in reports from off-wiki discussions
  • A participant pointed out that newer Wikipedians especially might encounter other Wikipedia users for the first time through a Discord or Telegram lens, and these spaces can be significantly less formal and rough around the edges in terms of topics, language used, form of messages, tone, etc.
  • There was a question about transparency: who can see who has been kicked or banned from the Discord channel, and a suggestion that it should be public
    • It was pointed out there is a log channel available to track actions that is accessible to authenticated admins and moderators
  • Suggestion: Clearer guidelines on not exporting on-wiki conflicts to 3rd-party platforms
    • Difficulty of defining conflict, where to draw the line?
  • Suggestion: Look at Technical Spaces Code of Conduct: Code of Conduct

Korean session[edit]

There were 14 Korean community members that participated this round table session. In the beginning of session, we discussed the behavioural issues regarding teenagers mentioned in the Diff blog, and then we discussed about proposals which were discussed during previous global round-table sessions in the May.

Topic 1 – Training to provide strategies to deal with cases of violation

Many participants agreed that there should be training for community members to be able to develop defence strategies against harassment.

  • User Training
    • Provide the manual to how to respond when you are harmed
    • There should be identification of what is considered unethical behaviours.
    • Develop storytelling strategies with the link
  • Admin / functionaries training.
    • Create a way to reach out/cooperate to the global enforcement body or foundation, so they are able to receive help when the violation of UCoC.
    • Share the cases from other languages, so that we can refer to these when necessary.
  • Who should be responsible for training process?
    • The local community, Foundation and Affiliates should all be responsible for providing/developing strategies against harassment or other kinds of attacks.
Topic 2 – Cross-wiki arbcom

We discussed the Ramzy's proposal about cross-wiki Arbcom, which was mentioned in the previous roundtable.

  • Should there be regional arbcom?
    • Ex) East-Asia (Korean, Japanese community)
    • Machine Translation Conversation, Inefficiency of Transmitting in English.
    • There are differences in the cultures between countries, so that it is difficult to make decisions accordingly.
    • However, since it is mediating the work of Wikipedia, it can not be regarded as a major problem. However, unless it is a technical problem, each culture is different, so it may be hard to review the cases.
    • Differences in cultural perception according to language (ex) East Sea vs Sea of Japan, Religious issue in South East Asia region. Like the Sincheonji, which is South Korea's pseudo-religion, foreigners have no awareness of what it is. So if a biased edit is happening and forward to regional arbcom, can it be solved?
    • There was an opinion looking this as another side of view: that it could be viewed as a means of looking at it objectively.
  • One Combined arbcom for Korean Language.
    • Since there is a lack of growth and active users in the Korean sister projects, no meaning for having the separate body.
Topic 3 – Private reporting pathways

During the LLC consultation, there was a demand to provide private reporting pathways, and there was an idea from Italian community to make a "report button". Like previous global round table sessions, there were concerns about implementing the private reporting pathways.

  • Disadvantages of private reporting: lack of transparency, Doubts about fairness of enforcement process
  • Report To Who? Local administrator, or new local body
  • If we implement the “report” button, there will be biased cases (Prank, misreport issue.)
  • Mailing list – A means by which threatened users can report, ex) We have already implemented this strategy for the case that is subject to record protection (by oversight).
  • Measures to protect users from attack are required ex) Fear of spreading as a means of retaliation, such as prohibiting sending emails using Special:EmailUser, and protecting user talks from harassment.

July 17[edit]

The third round-table session was held on 17 July 2021 with sessions planned for English, French, German, and Spanish speakers. The German session was held separately with French and Spanish hosted in the global room with interpretation support available. There were between 20 and 25 participants from over ten different languages and projects across both sessions.

This round-table session was planned for a longer length in order to also discuss the Movement Charter, Board elections, and other Movement Strategy initiatives. Input collected on these topics was provided to the respective teams.

Global session[edit]

The global session began with brief personal introductions followed by a short primer and update on the current stage of the Universal Code of Conduct project. Around 80 minutes were dedicated to this topic, including discussions about translation support, timing, and affiliate outreach for the upcoming enforcement draft guidelines review, how to apply the code practically, off-wiki or severe harassment, and private reporting mechanisms.

Upcoming draft review – translations and outreach

There was a question about language support, and a suggestion that community consultation stages should not officially start until translations have been completed.

  • This was a concern raised in the Phase 1 review, where some of the translations for other large projects came in later.
  • For the upcoming consultation, the Movement Strategy & Governance facilitators will be available to cover many major languages (though volunteer support will still be crucial).

There was a concern that 2 months would not be enough to have meaningful back and forth rounds, and a suggestion that the community should have an opportunity for revisions prior to moving to ratification (as it wouldn't make sense to reject something just because of a few issues).

  • The drafting committee timeline explains there is a refining period to follow the 2-month consultation so that the committee will have time to implement changes based on community comments.

Another suggestion was to try and get input from the readers of the project (e.g. non-contributors) in some way (some portions of the code do pertain to abuse of content).

Question: How will the consultation process be developed for the Affiliates, especially given the different status of a Chapter, Thematic Orgs, and User Groups? How will the Affiliates be able to provide input?

Answer: There will be outreach once the enforcement draft guidelines are ready for review. The Foundation’s facilitation team have meaningful conversations with the AffCom and Affiliates to determine how the guidelines will be integrated, especially for those who have their own CoC and/or internal behavioral enforcement mechanisms.
Application of the code

There was a question about whether the code would or should apply retroactively, to enforce behaviours in the past.

  • There are no enforcement guidelines yet but the general feeling was that this it would not be useful to do this.
  • There are existing terms of use, policies, and guidelines in place to guide current behaviour which can be relied upon.

A question was asked about how and whether local users will be expected to enforce or interpret the code, or who will perform this.

  • There was a concern that the code – read in certain ways – could constrain "regular wiki behaviour" (such as using off-wiki evidence to investigate unpaid editing, blocking someone for "competence is required", etc.)
    • These questions are being actively explored during phase 2
    • It was pointed out that, while the enforcement guidelines are not out yet, the code itself is meant to remain subject to appropriate context

There were some possible options brainstormed:

  • global team? (Volunteer or Foundation?)
  • local communities / existing bodies
  • tailored to the individual communities
    • E.g. Projects on mature enforcement systems: start locally, possibly with their arbcoms, for example: "UCoC button/wizard" that guides into existing community processes

Concern: If we offer too many options, we make all of them ineffective (never-ending appeals, the ability to appeal will reduce confidence in local governance structures).

There was a suggestion that the response needed for harassment is very different from the response to Vandalism. Vandalism suddenly becomes a UCOC violation, and it is done on a much more frequent scale. Does this vandalism go through the UCoC reporting system? (It was pointed out that vandalism is still often connected to harassment of community members anyway.)

Off-wiki / severe harassment

There was a question about whether any research had been done about possible unintended side effects: whether creating a UCoC on-project could lead to more harassment or poor behaviour off project. Social media channels were mentioned, which can often have a rather different atmosphere (more casual) rather than on-wiki (which most are not meant to be social networks).

  • One of the participants mentioned that due to some harassment, people are be encouraged to anonymize their usernames, and refrain from editing in a vast number of topics, to prevent people threatening violence, harassing phone calls, legal threats, and actual legal actions against contributors, feeling there were currently no effective protection.
    • In this way, harassment affects both the quality and neutrality of content and the user base.
    • Further, as projects start to lose their credibility or reputational value, the harassment seems to increase.
    • Sometimes harassment is even coming from non editors, government, or political bodies.
    • People are trying to "dox" some of the editors, to subject them to harassment or being accused of being part of a political party, and persecuting them.
    • Sometimes resolved by hiding usernames with revision deletion.
Private reporting mechanisms

It was pointed out that private reports can be in conflict with our ideals/goals about resolving disputes on Wikimedia.

  • Where do we draw the line on what can be a private report?
    • Different communities have different positions. Could this be decided by local communities?
  • Other questions about private reporting:
    • Is the Foundation requiring an effective private reporting mechanism from projects?
    • Is there a specific criteria of cases that will be eligible for private reports?
    • Will private reports always go to a local, or is there a way to forward them to a global body?
    • Will the reporting user have anonymity?

There was a concern that implementation of a private reporting system would increase the workload dramatically and overwhelm the existing bodies.

Participants were encouraged to engage with the draft review process, as these inquiries formed part of the key questions in the recent consultations, the results of which were provided to the drafting committee.

German session[edit]

The German session lasted for about 150 minutes with around 10 participants, spending time discussing several Movement Strategy and Governance topics. Christel Steigenberger gave an introduction on the Universal Code of Conduct, describing its origins and history, followed by an open discussion lasting around 80 minutes.

Details to follow.

Next steps[edit]

The input and ideas generated in these round-tables will be provided to the committee drafting enforcement guidelines. The facilitation team is continuing with additional round-table sessions with future dates to be announced. A comprehensive community review is scheduled in the project timeline for August to October 2021.