위키미디어 재단 커뮤니티 업무 위원회/2023-05-18 이사회와의 대화

위키미디어 재단 이사회의 위원회인 커뮤니티 업무 위원회는 5월 18일 18:00 - 19:30 UTC에 이사회와 대화를 주최했습니다. 이 대화는 커뮤니티 구성원이 자신의 업무에 대해 수탁자와 직접 이야기할 수 있는 기회였습니다. 이사회는 위키미디어 재단을 지도하고 그 책임을 보장하는 일을 담당하는 운동 지도자와 외부 전문가로 구성된 자원 봉사 단체입니다.

5월 18일 이사회와의 대화 영상 녹화본

참여 방법

이 대화는 줌에서 실시간 유튜브 스트림으로 진행되었습니다. 대화는 90분이었고 업데이트와 공개 Q&A 및 대화로 구성되었습니다.

대화 녹음은 위의 유튜브 링크에서 볼 수 있습니다. 공용에서도 곧 제공될 예정입니다.

질문 제출

참가자들은 공개 Q&A에서 실시간으로 질문할 질문을 가져왔고 미리 askcacwikimedia.org에 제출했습니다.

의제

• 재단의 연간 계획에 대한 커뮤니티 피드백; 재단 재정
• 위키축하

참고

Annual Plan and Foundation Financials

• This year’s draft plan has increased financial transparency
  • Fundraising models, revenue streams, budget breakdowns, department-level objectives, compensation and benefits principles, increased visibility for executive compensation
• Important step forward in commitment to openness and transparency
• Focused engagement on Product and Technology priorities, organized in three buckets
  • Wiki experiences: ensuring that we improve the experience of volunteer editors and editors with extended rights.
  • Signals and data services: Work to collect, store and provide analytics and machine learning capabilities for wiki experiences data and metadata.
  • Future audiences: Exploring ways of reaching new contributors and users.
• Published draft objectives and key results, discussed in community focus groups. That information is on the goals page of the Annual Plan draft.
• Discussions on buckets, on external trends in many places:
  • On-wiki: translated summaries of annual plan in 27 languages available on local language Wikipedias. Full draft translated into 10 languages on Meta-Wiki homepage. Engagement from community members on machine learning, technical collaboration, budget savings and layoffs, environmental footprint, work on equity, influence of China, language around describing different types of users.
  • Community calls: started in February at last Conversation with the Trustees. Call on AI in March, over 100 people attended and requested ongoing conversations on this topic. Joined 4 additional community calls where community members shared their plans with the Foundation. Talked about how to draft our plan with the rest of the movement. Engagement was around working with affiliates, addressing the gender gap, access to grants, support for copyright, IP blocks, how support will be divided and delivered to different projects.
  • Communtiy focus groups: 4 community focus groups on product and tech. Focus groups have talked about moderator workflows, IP blocks, and community technical support.
• Over 600 participants across community calls and focus groups.
• Translated feedback summary will be on Meta capturing major recommended changes. Will be sent to Board for approval in June and published in July.

Topic Q&A

Growth of executive compensation–is reducing it to historical levels preferable to layoffs?

• We build compensation levels based on data. The Foundation is a US-based org and people working at the Foundation need to be able to meet their needs.
• Compensation is not based on how much money we have year over year. We benchmark salaries to median base salaries (around the 50th percentile, corrected).
• Not feasible to return to what was offered years back. We are making data-driven decisions based on economies in the countries where we are hiring.
• Additional information in the Annual Plan and on Diff.

In reference to the Foundation’s new severance policy. Is the one month of severance pay based on the last month’s salary, the last year, or previous years? Will this policy affect severances for executives? For staff that are exceptions, do staff negotiate exceptions when they start, when they depart, or is it discussed during their tenure? How many staff are exceptions and will there be a maximum?

• Start with background. This was done as part of a holistic effort to look at our approach to staff across the Foundation; 50+ separate countries as well as many US jurisdictions. Harmonize where possible approach globally to severance. High level of complexity in harmonizing.
• Intentions for policy: communicate a standard to staff at all levels (including executives) the effort to put guidelines in place, given it is impossible to predict all situations. These guidelines will need to be reviewed regularly.
• One month is based on the last month of paid salary.
• The policy is too new to comment on what exceptions might be, likely they would relate to local laws and employment regulations. It could be unique circumstances that were difficult to imagine when we devise the policy. The purpose of the clause is to anticipate exceptions not to create a loophole.
• Contemplates natural limit of 9 months of pay.
• Intent to standardize practices where we can and continue to review.

Is there an incentive system in place to invite foundation staff to make donations to the Foundation or other wikimedia entities such as doubling their donations?

• Staff contribute to and participate in the movement in a range of ways. There is not a program in place currently but we do have staff that choose to be donors of their own accord.

Who approved these severance packages and when? Can we look at previous Board resolutions to find this information?

• The Board approved them. The Talent and Culture Committee is involved in these matters from the Board and makes recommendations to the whole Board.
• This involves HR and Legal questions. When there are resolutions like this related to personnel, they are not released publicly.
• The Foundation publishes everything it can publish, but resolutions of this kind are not published and are kept internal.

In this year’s Annual Plan, Maryana and Selena’s salaries were disclosed in advance. Is this type of disclosure something you’re planning to make regular practice?

• Executive compensation disclosures are required as part of the Form 990 and other forms.
• This type of disclosure may not be necessary in future years, but the intent is to continue to use the Annual Plan as a place to increase visibility, transparency and accountability of information from the Foundation.

Wikicelebrate

• Lightweight and easy way to celebrate Wikimedians each month.
• Focuses on Wikimedians who may not have been previously celebrated, especially long-term contributors.
• They receive a barnstar, dedicated page on Meta for celebration notes, Diff post, social media coverage.
• First recipient was Penny Richards: retired teacher that writes articles with others around the world. Major contributor to Women in Red and women’s biographies.
• Next recipient was Meenakshi Nandhini: Editor on Malayalam Wikipedia. Scientific articles and articles related to history and art.
• One goal for Wikicelebrate is to help connect the virtual to local events, so for May we will be celebrating and collaborating with Hackathon and EduWiki Conference organizers. You are welcome to nominate Wikimedians through an easy form!

Topic Q&A

How does Wikicelebrate work to complement what the movement has been doing to celebrate Wikimedians every year at Wikimedian of the Year?

• Reimagined Wikimedian of the Year three years ago. Created new categories: tech contributors, newcomers, laureate and more. This was received very well, allowed for new types of coverage.
• Last year we got great feedback: how can we celebrate more often?
• It’s simple, lightweight, and connect virtual to in-person events. It happens every month and welcomes nominations constantly. Allows us to keep celebrating each other and not miss out on wonderful contributions with everything else we have going on as a movement.

Open Q&A

How does the Foundation plan to work on AI/ML under this annual plan?

• Work around AI will fall under the Future Audiences bucket of this annual plan, which will represent 5% of our budget.
• The use of AI is expanding rapidly. Our content is already being used in ChatGPT. Goals will be about testing and learning because the change is happening so fast. We’ll need to learn a lot rather than making some large new investment immediately.
• The topic of AI inspires a lot of strong views, which makes sense because it is so consequential. We are actively looking for volunteers who are interested in being thought partners in shaping this work. If you’re interested email futureaudienceswikimedia.org.
• Machine learning has been part of our work for a long time. We have a Machine Learning team, even in the earliest days we had machine-assisted bots to help moderate content. That work fits under the Signals and Data Services bucket, and we’ll continue providing the same capabilities we’ve been providing to volunteers and staff. Things like suggested edits and prototypes to investigate ML powered interfaces that then support the work of the Future Audiences bucket.

Will Wikimedia have its own ChatGPT?

• If we did, it would be different. The opportunity to learn about large language models and whether training our own model fits into our world is interesting.
• Knowledge as a service–some of what we do works really well to feeding large language models. To do this well might require changes to our infrastructure. We want to look at whether the technology and infrastructure we have today support scaling–do we need to make that better and how?
• You may see more prototypes from us. Staff has already started working on exploring the space and testing the limits of what we have today.

What has changed in the Annual Plan as a result of community feedback?

• A comprehensive list of this will be shared on wiki once the feedback period is over.
• For example, some of the issues around moderation. We had targeted managing licensing issues that had grown in a queue. We proposed to community and they proposed a different approach and different goals. Shifted from improving an existing patroller workflow to looking at whether we can manage the intake of images and reduce the flow of images that need to be reviewed.

What is happening with grants under the next Annual Plan?

• Annual Plan document has a finance section, there is information there on grants. Headline is the grants budget is increasing even though Foundation budget is decreasing.
• There have been questions about specific grant programs on respective talk pages where people can find more information. (Funds overview here.)

What is the future of Movement Strategy work in the Foundation?

• Has to live in every part of the Foundation and affiliates.
• Last year, the Foundation began the process of centering our Annual Plan in Movement Strategy. This year’s plan builds on that by being more specific about the Movement Strategy recommendations and how to apply them internally across all our teams. Ensure it’s more embedded at the CEO level and in all aspects of the Annual Plan.

What is a “mutual separation agreement”?

• Many countries have governmental regulations that lay out terms for separations; the US has fewer of these than Europe, for example.
• Mutual separation agreements are a common feature of US workplace practices, it is exactly what it sounds like–when parties agree it is time to part ways, they then agree to what is expected of either party during the separation. When organizations and individuals decide together that they are parting ways.

Where is there more information about the Foundation’s finances beyond the 990 and Annual Plan?

• Financial reports come out every year
• Charity Navigator publishes information every year about the Foundation’s financials as well
• Diff posts published in the last months also discussed the Foundation’s finances

One of the goals in the draft Annual Plan does not have metrics yet: Effectiveness: Ensure our long-term sustainability by improving how the Foundation operates and scales. Can you speak about the process and timeline for this goal?

• For example, Charity Navigator assesses the amount of budget that is spent on programmatic expenses as a measure of effectiveness. In the 2021 year, the Foundation’s programmatic expenses increased by 4 percentage points to 77%. This can be used on an ongoing basis to measure effectiveness
• Each department will work toward their own metrics but share them across the foundation to ensure work is not siloed. That the intent of making this Foundation-level objective.

Can you talk about how moderator workflows will be integrated into our work in the upcoming year?

• It’s a combination of work that was already in progress and problems that were already identified (for example IP masking, we need to change the public visibility of IPs).
• That has some consequences in terms of the tooling. Example of an area where we’ve been collaborating for a long time with communities.
• Another area: community configuration. There are a number of features that we’ve seen through testing increase the number of new editors that come in. When we increase the number of new editors, that increases the burden on existing editors to review and manage the content. Giving communities the opportunity to make decisions about these features to local wikis is something we’re looking at.
• Abuse filters are another one. Are there ways we can augment that with additional machine learning models or UX improvements?

What you’ll see in the updated plan are statements about the outcome we’re trying to achieve with the work of a team or larger group. Attaching metrics or qualitative review to assess impact.
What is the plan for determining how to divide resources in the next year among different Wikimedia projects?

• Will have a more complete picture once the feedback cycle finishes and is reviewed. We will continue to invest in all language Wikipedias and do infrastructure work that improves MediaWiki. This will affect every project. There are also infrastructure projects around the way that we deploy the software which will also improve experience on every project.
• Investment in Commons–a dedicated team.
• Goal this year was to get into a starting position and be able to talk more about investment in data and infrastructure, and improving the core user experience in line with movement strategy.
• Once we process all the feedback received we’ll have more information about how resources are to be divided project-wise.

Can Selena and Maryana talk about the principles that guided your work on the Annual Plan and what you hoped to bring to the process?

• Talking about what our mission is and what we’re here to do was critical. Two things top of mind on product and tech front:
  • Supporting our volunteers who produce free knowledge and are we serving their needs. There were clear gaps.
  • Looking at the infrastructure we’re using to share this knowledge–is that healthy? How do we determine whether it’s healthy or not? We have to start measuring.
• Having really clear deadlines and being flexible at certain points. That was the work of many people making sure that we were clear about dates but also giving people enough time to make big decisions and think about our direction.
• There was two-way planning and also a healthy back and forth between management and individual contributors being able to think about and propose work.
• Aligning the Foundation’s planning directly to Movement Strategy has helped a lot to build off the past.
• Clarifying our metrics which is ongoing.
• Taking the principle of two-way planning seriously. Ensure that the Foundation is planning against the work that is being done by affiliates and communities across the movement.
• Bringing multilingualism much more forcefully into our planning work. Having the plan available in 27 languages feels like momentum in recognizing this.
• There’s so much internal work happening with a One Foundation approach to not have different teams feel they’re trying to manage and plan on their own. More integrated.