Talk:Wikimedia Foundation Annual Plan/2024-2025

Draft plan feedback

edit

Tech grants

edit

Will this plan include the reintroduction of grants for technical projects not fundable under the rapid grants program? Such a program was available before the grants refresh, was promised after the grants refresh, but never materialized. AntiCompositeNumber (talk) 01:07, 13 April 2024 (UTC)Reply

Thanks for this question, @AntiCompositeNumber, and I’m sorry it took a while to provide an update on this. Over the past year or so, we’ve been working with the Product & Tech department to better align grantmaking to technical projects. In the past, we recognized that we would from time to time fund projects that seemed like a good idea, but that were not sufficiently aligned with WMF’s Product & Tech roadmap. This would lead to very understandable community frustration about the plethora of unmaintained tech in the ecosystem.
As you noted, when we launched the grants refresh strategy, we put the Tech Fund on hold so we could ensure that when we started funding technical work again, it was in alignment with Product & Tech. Since then, WMF hired a new CPTO, so we further put this work on hold to allow her and her leadership team time to align on their strategy.
That brings us to today, and I now have an update on this work. We will also update the relevant page on the Community Resources page on Meta.
We will not have a dedicated Tech Fund in the near future. Instead, if community tooling is needed, the Product & Tech department will work directly with volunteers to develop a plan for building it (potentially using contract agreements vs. grants), alongside a plan for maintaining it.
Relatedly, as part of the changes being made to the Community Wishlist Survey, the WMF Comm-Tech team intends to invite and support volunteer developers to join in the effort to help resolve technical issues raised by community members. The broad objectives are aimed at improving participation in the wishlist survey process towards wish submission, and wish resolution (details). In March 2024, WMF conducted the first Community Wishathon - a virtual event for community developers and WMF staff to work on technical issues submitted through the recent wishlist surveys. From the later half of 2024, the community wishlist survey will remain open permanently as an intake mechanism for technical wishes. As a result, we will introduce changes to address those wishes with better engagement and collaboration. Besides periodic events like community wishathons, we would like to continuously support volunteer developers willing to take up some of these technical issues by providing guidance, reviews, mentorship, and assistance with long-term maintenance plans for that software. If the need arises for specific work to be done via contractual agreements, we will explore those opportunities as well. RWeissburg (WMF) (talk) 13:28, 16 April 2024 (UTC)Reply
Got it. The work that would have been done will continue to not be done. AntiCompositeNumber (talk) 13:29, 16 April 2024 (UTC)Reply
Well said. In the absence of a well-articulated, well-supported infrastructure, little is likely to improve. Mary Mark Ockerbloom (talk) 02:33, 29 May 2024 (UTC)Reply
So if I understand correctly, WMF is planning to enhance the collaboration between WMF themselves and volunteer developers, and funding / contractual agreements will become a part of the collaboration and support? I am concerned about if developing maintenance plan (e.g. setting up the schedule) will affect the motivation of volunteers as it will limit their flexibility. Thanks. SCP-2000 15:56, 16 April 2024 (UTC)Reply
Getting hired/contracted by the Foundation is a lot higher bar than grant-funded work, and it absolutely won't work for previously grant-funded work including the building and maintenance of community tools. AntiCompositeNumber (talk) 20:20, 16 April 2024 (UTC)Reply
This will inevitably bring much less attention to the problems and wishes of non-English speaking communities. Tech grants allowed people to start developing tools that didn’t need a buy-in from English Wikipedians, making everything depend on CWS only re-inforces the systemic biases of both the WMF and the Wikimedia movement at large. stjn[ru] 14:46, 17 April 2024 (UTC)Reply
Hello @Stjn. The upcoming changes to the community wishlist survey are intended to expand the visibility into technical needs of non-English communities, which have historically not participated in the wishlist much. There are several developer volunteers working individually or as user groups (formal and informal) across the world who have greater visibility into the technical issues faced by the communities around them. We plan to reach out to these groups and individuals to better coordinate about technical needs, and understand how these issues can be addressed. This will also help us build a better connected network of technically competent people who can work alongside the WMF Product and Tech teams. Some of these collaborations are already happening, and we are promoting them further through events like the Community Wishathon, or the Wikimedia Technical Summit in India. Runa Bhattacharjee (WMF) (talk) 03:22, 24 April 2024 (UTC)Reply
This approach gives me pause. Without well-advertised, open calls for participation, a plan to "work directly with volunteers to develop a plan" will typically revert to informal word-of-mouth social networks that have problems of their own. My hope is that there are more structured engagements that are planned.
I would feel more confident if we had better connections from the communities to WMF staff and processes. However, it has always been hard for the communities to understand or track who are the right entities to be interacting with or what projects are still functional.
For example:
  • The current public staff/contractors chart does not reflect the fact that the VP of Product Design Margeigh Novotny left in September 2023.
  • There is currently a Wikimedia Foundation "Future Audiences" group (setup in May 2023?) but it does not show up on that chart.
  • Sometimes an email bouncing, mid-exchange with WMF staff, will be the only indicator that the person has left or that a department has changed. This was the case recently with a Knowledge Equity Grants project and a WMF VP.
These cases may be minor, but they do point to the challenges of staying in touch especially since the staff/contractors list is not on a wiki anymore so tracking these changes is increasingly difficult. - Fuzheado (talk) 15:25, 23 May 2024 (UTC)Reply
I created phab:T344479 in 2023 to try to get the staff and contractors page updated. Feel free to make some noise in that ticket. However, I think that WMF has a private OfficeWiki where they store this information, so they have little incentive to update the public page. –Novem Linguae (talk) 23:47, 23 May 2024 (UTC)Reply
Hi Fuzheado and Novem Linguae, I wanted to comment on knowing who is working on what. It is a good point, and although there were owners listed for the work around the product & tech goals, this information wasn't available for the other goals. We will make edits shortly to the annual plan to be clear who's the best person to contact for which part of the annual plan.
Also, on the more general point about the staff and contractors listing, we have been working over the last few months to move the listing on to Meta-Wiki - as you rightly note the Foundation website is out of date. The website will be updated to match Meta, but Meta will serve as the main listing going forward as part of a larger body of work. We've just posted an update on this work on the relevant project page--you'll see the work is part of a project to better connect community members with the information they’re looking for from the Foundation. In addition to building staff listings on Meta and establishing a process to keep them up to date, this work has also involved a clear, visible, human powered way to contact us on any topic. So if a staff member listed here moves on, or if you don’t know how to reach out about a particular topic, you will be able to email answers@ to get your question triaged and answered. ELappen (WMF) (talk) 20:10, 30 May 2024 (UTC)Reply
@ELappen (WMF). Awesome. That all sounds great. Thanks for your team's work on this this week (lots of good progress) and for your future plans. –Novem Linguae (talk) 13:07, 31 May 2024 (UTC)Reply
Hi Fuzheado - The improvements to the Community Wishlist that I mentioned and Runa elaborated on are intended to be exactly what you name as needed, "a well-advertised, open call for participation." If you have suggestions for how to better advertise it, or ensure more people participate in it, let us know! RWeissburg (WMF) (talk) 23:39, 31 May 2024 (UTC)Reply

Hi AntiCompositeNumber, I'm glad someone already asked this question.

Some quick thoughts: if we group past technical projects and proposals by scale, we might have

  • Strategic projects needing > $3M to get off the ground, with extensive ground work needed before their core use materializes (Wikidata, Abstract WP). Generally funded or co-funded by WMF with plans integrated into shared long-term technical and content visions
  • Projects needing $300k-$3M to get off the ground (Wikipedia itself, Commons, the collected Manskeverse, Wikia, SDC, Shared Citations❌, DPLA integration into Commons). Funded and maintained in a diversity of ways – self-funded, crowd-funded, vc'ized, targeted grants, never funded. Initial funds generally give rise to further ones.
  • Projects needing $30-$300k (TranslateWiki, dbpedia ←→ wikidata fact sync, Balinese palm leaf transcription integration w/ Wikisource (1 + 2), Migrating the Lingua Libre project to a MW+Wikibase instance, OpenRefine support for Commons batch jobs, developing a Wikibase loader for a popular data pipeline tool, AI category suggestions for Commons images❌, making Timeless a baseline for future skin design❌). These often streamline existing community-maintained work, make it accessible to new audiences, or port things to be compatible with existing frameworks.
  • Projects needing $3-30k (Global watchlist, The Wikipedia Adventure, making Timeless easy to maintain, WD editor for WP❌, VideoWiki dev support (1,2,3❌)). These have been more diverse in scope and impact, sometimes (often?) intended to be maintained by the creators or their community of non-technical curators.

Only a tiny subset of these ideas make it to the top-5 of a community-wishlist. Wishlist ideas may or may not have a clear implementation; a technical proposal has both implementation and willing implementers with time. I'm not sure how much of the broad-ranging community work would be considered in scope for the core team, focused on a small number of big-ticket concepts. But the synching, scripting, cleanup, and merging-in of other small communities of [tools and practice] is part of the living ecosystem of the wikis. And for the most part, we keep our tools and core separate...

RWeissburg thanks warmly for your comments. Some related questions:

  1. The current roadmap [at least WMF's roadmap; maybe the rest of the wikiverse needs to merge and articulate its joint roadmaps?] doesn't say much about content-facing or user-facing products; most of which are created and maintained by individuals + other groups in the movement. Are things not explicitly on the roadmap by default considered 'aligned' or 'unaligned'?
  2. How can smaller community scripts and tools get grant support from within the wikiverse, unless it is through regional affiliates? a) if directly related to something on the roadmap; b) if not?
  3. How can substantive integrations, simplifications, and large-scale content improvements get support? given that our formal Plans and roadmaps are fairly silent on the topic of reaching out to new knowledge communities or scaling up our inclusion of knowledge from existing partner institutions?
  4. How should larger projects (think WikiCite) think about getting support, say from a mix of wiki and matching state or external-foundation funding?

SJ talk  23:39, 21 May 2024 (UTC)Reply

Hi SJ,
I don't have answers for most of your questions, though I think they're important ones. What we're trying to address by not standing up the Technology Fund is ensuring that we don't further exacerbate an issue I know many Wikimedians feel acutely - that in the past community tech has been funded, grown to be depended on by many volunteers, and then gone unmaintained, because the volunteer(s) who built it were no longer able or interested in maintaining it, it became obsolete as WMF released new tech that addressed similar issues, it became too expensive to maintain as it was built, or something else.
We're trying to address this by the strategies I mentioned above: improving the Community Tech Wishlist as @Runab WMF noted, and, when appropriate, asking volunteers to partner with WMF's Product team more strategically (through the aforementioned contracting process) to plan out the build and maintenance. In those ways we're hoping that whatever is built by community members with funding from WMF / community funds, is aligned with the overall product strategy.
I'm curious how you would think about addressing this? There's certainly the option you mentioned in #4 about going to external funding sources, but I worry that the same issues might arise.
I'm open to suggestions or recommendations you have, and imagine @Selena Deckelmann is as well (also ccing @Guillaume (WMF) here for line of sight). RWeissburg (WMF) (talk) 00:37, 30 May 2024 (UTC)Reply

Financial Model

edit

What do you mean by "Invest in longer-term strategies that will provide increased revenue in a future where Wikipedia content is primarily consumed off our platform."? --Mervat (talk) 02:27, 16 April 2024 (UTC)Reply

Hi @Mervat, as we see in the external trends section of the plan, the ways people are searching for content online and the type of content they’re searching for is changing. People are increasingly consuming Wikipedia content off our platform, which poses challenges to the banner fundraising model that has been the primary revenue stream historically. The changing online environment and a future where readers may no longer primarily read content directly on Wikipedia requires us to invest in longer-term strategies that are more resilient to changing readership trends: Wikimedia Enterprise, the Wikimedia Endowment, and a growing Major Gifts program. We discuss these revenue streams in the long-term revenue strategy section of the plan. The movement strategy revenue working group also discussed the need for these long-term strategies as knowledge becomes more granular, remixed by others, and served through other interfaces. MeganHernandez (WMF) (talk) 21:06, 17 April 2024 (UTC)Reply
Thank you @MeganHernandez (WMF) --82.212.106.8 (talk) 14:14, 23 April 2024 (UTC)Reply

More technical support for Wikimedia Commons

edit

Hello – for certain reasons, my litmus test for this annual plan was if and how it budgets more technical support and development work for Wikimedia Commons. Can I ask you to point me to the relevant section? Thank you, Gnom (talk) 20:00, 18 April 2024 (UTC)Reply

Hey @Gnom, while it's not explicit in the text, some support for Wikimedia Commons is planned as part of Objective & Key Result WE2.3. The implication of this are still being defined by the people who will be in charge of this objective, so I can't go into detail, but there will be some support and development work going around Commons also for next fiscal year. Hope this helps! Sannita (WMF) (talk) 11:44, 2 May 2024 (UTC)Reply
@Sannita (WMF): Thank you – but this leaves me confused. This means that zero work on Commons is in the annual plan? Gnom (talk) 12:36, 2 May 2024 (UTC)Reply
@Gnom I cannot still confirm how work on Commons will play out, but there will be some level of support and bugfixing for Commons. Sannita (WMF) (talk) 12:39, 2 May 2024 (UTC)Reply
Well, we need to underline that there is zero mention of any such work in the current draft of annual plan. Commons desparately needs more development resources – not just "some level of support and bugfixing" – and the annual plan should absolutely reflect this. Gnom (talk) 07:27, 3 May 2024 (UTC)Reply
You're not the only one suggesting this, and I will report that there is cross-wiki substantial consensus to get more info about this kind of data. Sannita (WMF) (talk) 15:40, 3 May 2024 (UTC)Reply
Yes, please. —Ismael Olea (talk) 15:18, 28 May 2024 (UTC)Reply
This is so, so, so funny. Money for the UCoC, that nobody wanted, to supress the Communities is here. And a loooooot of money for the paid staff. But for the Communities and the projects: nothing.
"WMF: We're treating our volunteers like s**t since 2004" would give a great begging slogan for the next fundraising campaign. And it would be for the first time true. Marcus Cyron (talk) 23:17, 31 May 2024 (UTC)Reply
I would like to +100 this comment from @Gnom. I fear this annual plan has only "Infrastructure" in the list of high level goals, and does not specify anything substantial about our interaction and relationship with the audience in terms of multimedia or rich content beyond the text of Wikipedia. I worry greatly about the specific emphasis on this: "Improve User Experience on the wikis, especially for established editors and functionaries." What about public-facing experiences, and evolving our platform towards visual learners, the Tiktok generation, those on mobiles, and more? I recall one of my earliest conversations with SDeckelmann-WMF was in Berlin (September 2022) about her concern on how we might be relevant in the era of Tiktok. Therefore, it is a surprise this doesn't show up at all in the annual plan.
While Wikimedia_Foundation_Annual_Plan/2024-2025/Goals/Infrastructure mentions MediaWiki specifically, it does not mention Wikimedia Commons, nor does it mention interactive, multimedia, video, or Wikidata. To its credit, it does mention reviving the defunct Graph extension sometime in the future. But this would be an ideal time to also allocate resources or strategic bandwidth to Think Big (see the 2022 letter) about not just the Graph extension, but about a whole slew of backlogged requests in Phabricator related to Commons, such as support for 3D file formats (Phabricator), VR/360 photography (Phabricator), video , and more.
It is times like this that I believe we made a mistake in the Movement Strategy. We stated too emphatically about being the "essential infrastructure of the ecosystem of free knowledge," without also mentioning user-facing aspirations and priorities. I fear that we are taking it way too literally to the detriment of evolving our approach to content and what is actually contained in that infrastructure. Since the establishment of that "infrastructure" phrasing, we have seen the decline or under-resourcing of Structured Data on Commons, GLAM Wiki tools (see GLAM manifesto or GLAM CSI), and Commons in general (see previous Think Big letter). If there is still time, I would urgently request the addition of more audience-facing thinking, as right now it feels like a significant oversight. Tag: Ziko, Frank SchulenburgFuzheado (talk) 17:13, 21 May 2024 (UTC)Reply
@Sannita (WMF): For what it's worth: I'm an admin on Commons, active for 20+ years in Wikimedia projects, and I can only fully support the statements by Gnom and Fuzheado. Commons is barely clattering along on its rusty mechanical legs unto which the community has heroically bolted all kinds of rickety support structures and cogs just to keep it more or less usable, but these tools break all the time, not to speak of meaningful video support or advanced features for using Commons content. Gestumblindi (talk) 19:18, 21 May 2024 (UTC)Reply
@Sannita (WMF): as yet another longer term contributor at around 19 years and creator of Commons biggest(my measurement) project; Quality Images Commons has really been left to stagnat, even wither on the vine. This plan doesnt look to bring Commons even up to a reliable service level, there is nothing that indicates any concept of far media formats have progressed since its inception. While it may be seen as backwater the WMF doesnt want to invest in, until Commons is improved and advanced every other project is hamstrung by it. Gnangarra (talk) 15:09, 22 May 2024 (UTC)Reply
I think the apparent disregard for the needs of Commons shows a lack of understanding of Commons' dual importance: First, its fundamental role as a supporting project for most Wikimedia projects, a central media repository that just has to work reliably. Second, Wikimedia Commons as the only place on the internet with a community-curated collection of truly free educational media content that fulfils at least CC-BY-SA standards, that is, a project which has the intention to guarantee that you can freely reuse (even commercially) and modify the content (I say "the intention" because the community is often overwhelmed and there's a lot of content on Commons that shouldn't be there; also something for which we would need more and better tools to manage). Other platforms like flickr (for images) or commercial giants such as YouTube are much, much larger, but none of these have Commons' vision of scope, quality and licensing standards, all handled by a rather small, but passionate community. Gestumblindi (talk) 20:03, 22 May 2024 (UTC)Reply
Dear colleagues @Fuzheado, @Gestumblindi, @Gnangarra, @Gnom: I can only agree with your assessment. It is obvious what priority the WMF gives to our Wikimedia Commons media collection. A very low one. In any case, it looks like the opposite of "Think Big".
Perhaps the WMF is thinking: WMC works somehow, and if necessary the volunteers will just have to invest more work to upload something and make the images accessible. It won't be the intention, but you get the feeling that volunteer work is seen as an infinite resource.
And what about the cooperation with GLAM institutions? What is the message to them? I wouldn't even want to elaborate on that here. Ziko (talk) 10:02, 25 May 2024 (UTC)Reply
Dear all. I'm reading your messages here and I couldn't agree more. I have tried to explain this to @MPinchuk (WMF) in this other discussion, with very poor results. We urgently need improvement for Commons, and this can't be limited to trying to solve some bugs. We could be the only educative video platform without ads in the web. Instead, the Annual Plan tries to add a button to Flickr so people can share images. A button that won't be added by Flickr, by the way (you can read the details in the discussion). Theklan (talk) 15:25, 29 May 2024 (UTC)Reply
Lots of opportunity for growing Commons with some further technical infrastructure. For example we right now have no way of knowing how often videos are viewed. This makes it hard to push further video development if we do not know how often this material is viewed. Doc James (talk · contribs · email) 19:41, 29 May 2024 (UTC)Reply
I can only second the statements of Gnom, Fuzheado, Gestumblindi and others. We all put thousands of hours of work into keeping Commons up and running, but it’s often like fighting windmills – there is such an incredible amount of work to be done and our tools break all the time ... Dear Foundation, you are ruining your most important capital: the committed volunteers. – Aristeas (talk) 08:04, 9 June 2024 (UTC)Reply
@Doc James: mediaviews, although I don't fully know how it is counted. Best, —DerHexer (Talk) 14:18, 20 June 2024 (UTC)Reply
It counts the number of pageviews for the pages that the video is on. It does not count the number of times the play button is clicked unfortunately. At least that is my understanding. Doc James (talk · contribs · email) 14:34, 20 June 2024 (UTC)Reply
@Doc James: I'm not sure … But I also mixed up the tools. I wanted to link to https://mvc.toolforge.org/ which was created by Ladsgroup. But this relies on mediaviews so that the same question remains open. Best, —DerHexer (Talk) 12:39, 3 July 2024 (UTC)Reply
Thanks. We are looking at the possibly of funding improvements around video views in our budget next year. Will reach out to Ladsgroup. Doc James (talk · contribs · email) 12:43, 3 July 2024 (UTC)Reply
Addendum: The page Wikimedia Foundation Annual Plan/2024-2025/Product & Technology OKRs outlines the "objectives and key results" for the year's work portfolio "buckets." The following breakdown of the frequency of specific named projects is an imperfect indicator, but it may help illustrate the issues described above.
Wikipedia mentions: 13
Wikifunctions mentions: 3
Wikidata mentions: 1
Wikisource mentions: 1
Wikimedia Commons mentions: 0
Fuzheado (talk) 17:02, 30 May 2024 (UTC)Reply
i just think Commons has such incredible future potential, that to under-resource it's development now is to kick ourselves in the face (which is bad!) Lajmmoore (talk) 16:11, 31 May 2024 (UTC)Reply

Among the many areas which need care: Commons:CropTool is broken, Commons:video2commons is broken. Yann (talk) 19:24, 31 May 2024 (UTC)Reply

Knowledge diversification

edit

Commons is an important project that deserves more love (and not just from the WMF). However, maybe I could broaden the question a bit beyond Commons: as a movement, we're very dependent on Wikipedia. I see some value in diversifying our movement with regards to how we develop, aggregate, present and share knowledge. Could you perhaps talk about if and how the current Annual Plan actively supports that diversification (e.g. support sister projects more actively than before, or provide more infrastructural support for volunteers to do so)? Effeietsanders (talk) 04:48, 23 May 2024 (UTC)Reply

Hi @Effeietsanders,
Thanks for your call to broaden the strategic question that is in front of us with regard to diversification of knowledge. There are many ways to think about diversity, and I’ll take on two different lenses to answer – there are definitely more ways to think about this and I encourage others to weigh in.
One way to think about diversity when it comes to development and aggregation is the scope and scale of the content and subject matter that is covered in our ecosystem. Within this, we have invested in understanding knowledge gaps, through extensive research and development of taxonomies, which the Foundation and many participants in the broad movement have developed together. This research has shown us many areas where we have gaps (notably regarding gender and geography, but of course in more areas). We continue to invest in addressing knowledge gaps in a variety of ways, for example by developing and supporting systems to enable volunteer-created lists of content to be developed, and continuing to expand knowledge creation capacity through translation technology and support for training of volunteers in existing and new languages. This kind of work is definitely in the annual plan.
Another way to support diversity when it comes to presenting and sharing knowledge is through the Movement Strategy vision which calls out “knowledge as a service” as a key element in our future. We have implemented this through our APIs, and an [“enterprise” API system] and related products. To meet the challenge of our mission to disseminate knowledge worldwide, we need to focus on the user experiences that meet a need people have. A measurement of meeting that need is the volume of pageviews and unique devices, which are quite large for Wikipedia, and are increasing in certain markets. In addition, we’re seeing adoption of the Enterprise APIs, and increasing automated traffic (from a variety of sources) from what we believe are companies who wish to use our freely licensed data in their products. Improving our APIs enables broad access to the knowledge that has been assembled in many projects, and has supported research, products and integration with the whole of the internet. In next year’s annual plan, we are continuing to deepen our investment and maintenance work for MediaWiki itself, which is the beating heart of all of our projects. By improving the state of the APIs supported by MediaWiki and related data services (including WikiData and Commons), we will improve the world’s ability to quickly and reliably access, use and share data that’s part of our systems.
All that said, Wikipedia itself and other Wikimedia projects will continue to be destinations and need support for web and mobile experiences for the foreseeable future. But we must continue to also invest in APIs to realize the Movement Strategy goals. When it comes to the Sister projects, we have a board committee that includes volunteers that is addressing those concerns and is currently seeking feedback on their work.
A key question for me is: is there a clear demand for knowledge diversification in the form of different not-for-profit user experiences (specifically: not Wikipedia and not encyclopedic articles)? I have seen for profit use cases, primarily to aid in selling advertising online. I would love to hear from anyone who has knowledge and data about this. SDeckelmann-WMF (talk) 00:41, 6 June 2024 (UTC)Reply

Support for Commons is Revenue generating

edit

The first strategy of this annual plan is "FINANCIAL MODEL: Wikimedia's financial model and future projections for revenue streams in online fundraising".

I object to the Wikimedia Foundation's disregard for Wikimedia investments to content. The WMF only counts revenue and financial donations when those go to the Wikimedia Foundation, but does not make an effort to record, recognize, or thank donor organizations who spend large amounts of money in developing content for Wikimedia projects.

I have been a Wikimedian in Residence since 2012. In that time my own employer institutions have invested money in me, my outreach, and a lot of their partner organizations to develop Wikimedia content. I do not need or want direct recognition from the Wikimedia Foundation, but I and all the other community members who persuade organizations to make content donations to Wikimedia Commons would appreciate some basic infrastructure to justify why contributing to Wikimedia projects is a good idea. There was a time recently in New York City when several museums and universities had programs which cost them hundreds of thousands of dollars a year to do content development in Wikimedia platforms, and the only thing they wanted in return was basic digital metrics to verify that content was posted and that people could access it. Unlike all other user generated content platforms, the Wikimedia Foundation has never been serious about providing professional data which is the baseline for enabling partnerships. Those partnerships have gone - perhaps with major institutions reporting dissatisfaction with the Wikimedia Foundation - due to unreliability of tools and inability to get basic impact metrics.

Back in about 2012 the WMF decided on some metrics without community input and in 2014 started mandating them in Learning and Evaluation/Global metrics and various legacy systems. The WMF strategy emphasizes investments which result in large numbers of edits in low-traffic places, and de-emphasizes development of high-traffic content. This has always been a strange strategy and never based on community consultation, and has always been weird compared to what content creators do on other social media platforms. There is no successful YouTuber, Instagrammer, etc who makes a conscious effort to produce huge amounts of low traffic content; instead they all want popularity and attention. Every platform other than Wikimedia encourages content creators to develop content for high-traffic places to meet reader demand. This is what our partner organizations want, but not part of the WMF strategy.

GLAM and Commons are unable to focus on popular areas because we have no way of getting proper impact metrics. For lack of development time 10 years ago and every year since, we are continually losing out on wealthy partner museums, universities, and other knowledge centers who would be using their own money to organize their own wiki programs, if only we could give them some digital metrics which indicates that their development of Wikimedia Commons has an impact to users.

If there is a goal to get new revenue into the Wikimedia Movement 1) recognize and appreciate that many large donors give money by investing directly into Wikimedia content and 2) give those donors the measurement tools they need to justify to their own stakeholders why they should post to Wikimedia instead of elsewhere. Bluerasberry (talk) 01:40, 29 May 2024 (UTC)Reply

  Support —Ismael Olea (talk) 09:11, 31 May 2024 (UTC)Reply
  Support Yann (talk) 19:25, 31 May 2024 (UTC)Reply
  Support Gnangarra (talk) 10:09, 1 June 2024 (UTC) the less WMF invests in Commons, the more others will take from the movement projects to offer enriched multimedia experiences, resulting in less contributors and funding.Reply
Thanks for sharing this perspective @Bluerasberry. I agree that it is vital that GLAMs understand the impact of their contributions to our projects. That’s why Commons Impact Metrics was prioritized in last year’s annual plan and we’ve been reporting progress on the WMF support for Commons page each month. We’ll have a more substantial update by the end of June.
I’m interested in your proposal to recognize institutions that make significant content ‘donations’ to our projects, or host Wikimedians in Residence. Do you envisage this as a list of contributing institutions? I know there is already an approach for documenting Wikimedians in Residence on Wikidata. Are you aware of other efforts like this? I am a member of the working group convened by Evelin Heidel to redesign the public presence of GLAM Wiki on Meta and will see how we can take this idea forward. FRomeo (WMF) (talk) 11:02, 6 June 2024 (UTC)Reply

Resourcing Wikimedia Commons beyond Wikipedia

edit

The below statement has also been published on Wikimedia Commons. Endorsements here or there (or both) are very welcome.

On Wikimedia Commons, the Wikimedia community works together with educational and cultural partners around the world, collecting images, videos and other media in alignment with Wikimedia's vision (the sum of all knowledge), and Wikimedia's 2030 movement strategy's strategic direction (infrastructure for free knowledge, knowledge equity). In a world where much knowledge has never been written down and in a time where people around the world increasingly communicate and learn via audiovisual means, we are dedicated to connect, synchronize, contextualize, improve, curate and share a "multimedia knowledge commons". We collect and describe humanity's knowledge as captured in images and video, in art and photography, in digitized publications and recorded culture, and so much more. The 100 million files currently on Wikimedia Commons are only a start.

We do this because images, videos and other media convey knowledge in unique ways. They explain and demonstrate complex concepts, cultural expressions and historical moments across languages. They do so with clarity and nuance that words often can't achieve.

Therefore, Wikimedia Commons' scope is, and must continue to be, much broader than just support and illustration for Wikipedia. So much of the world's knowledge is not, and may never be, properly describable in and/or notable for Wikipedia. Wikimedia Commons is necessary for a movement that aims to collect and share all the world's knowledge; we need it as a platform that allows us to transcend Wikipedia's text-based limitations.

The functionality of Wikimedia Commons, with this broad scope, must be properly resourced by the Wikimedia movement and Wikimedia organizations, so that Wikimedia Commons can reach its full impact, enabling and encouraging diverse paths of discovery, active participation, and broad re-use.

In the context of the Wikimedia Foundation’s 2024-25 annual plan, we ask:

  1. Wikimedia Foundation staff, teams and leadership should be aligned to fully support Wikimedia Commons with its broad scope as a key piece of mission- and strategy-aligned knowledge infrastructure in its own right.
  2. Wikimedia Commons, with this broad mission-aligned scope, should receive proper resourcing in the 2024-25, and future, WMF annual plans.
  Support Spinster (talk) 07:27, 31 May 2024 (UTC)Reply
  Support —Ismael Olea (talk) 09:09, 31 May 2024 (UTC)Reply
  Support Susanna Ånäs (Susannaanas) 🦜 11:42, 31 May 2024 (UTC)Reply
  Support Scann (talk) 14:28, 31 May 2024 (UTC)Reply
  Support Fuzheado (talk) 18:03, 31 May 2024 (UTC)Reply
  Support GPSLeo (talk) 19:01, 31 May 2024 (UTC)Reply
  Support Yann (talk) 19:16, 31 May 2024 (UTC)Reply
  Support Jmabel (talk) 19:16, 31 May 2024 (UTC)Reply
  Support Ambrosia10 (talk) 20:08, 31 May 2024 (UTC)Reply
  Supportsigned on Commons too Gnangarra (talk) 10:05, 1 June 2024 (UTC) -Reply
  Support MartinPoulter (talk) 12:30, 3 June 2024 (UTC)Reply
  Support --Don-vip (talk) 17:50, 3 June 2024 (UTC)Reply
  Support -- Bodhisattwa (talk) 17:57, 3 June 2024 (UTC)Reply
  Support --Ciell (talk) 19:48, 3 June 2024 (UTC)Reply
  Support --Prototyperspective (talk) 22:02, 3 June 2024 (UTC)Reply
  Support -- JWilz12345 (Talk|Contrib's.) 23:35, 3 June 2024 (UTC)Reply
  SupportAristeas (talk) 09:22, 4 June 2024 (UTC)Reply
  Support -- Radomianin (talk) 09:55, 4 June 2024 (UTC)Reply
  Support -- Giles Laurent (talk) 12:59, 4 June 2024 (UTC)Reply
  Support --A.Savin (talk) 20:56, 4 June 2024 (UTC)Reply
  Support but what I said in the duplicate thread on CommonsRhododendrites talk \\ 01:43, 5 June 2024 (UTC)Reply
  Support Suyash Dwivedi (talk) 16:37, 7 June 2024 (UTC) (on behalf of CPUG)Reply
  Support --Herbert Ortner (talk) 18:32, 9 June 2024 (UTC)Reply

Response

edit

Point of information – a response from Selena Deckelmann was posted here: commons:Commons_talk:Media_knowledge_beyond_Wikipedia#Response_from_the_Selena_Deckelmann,_CPTO_at_the_Wikimedia_Foundation - Fuzheado (talk) 15:11, 6 June 2024 (UTC)Reply

Global Majority Wikimedia Technology Priorities

edit

I am writing to bring to the attention of the Wikimedia Foundation (SDeckelmann-WMF and others) and the group following discussions on this Annual Plan a recently released document called Global Majority Wikimedia Technology Priorities. Points this document makes are relevant for the WMF Annual Plan.

This document lists ten priority points for the Wikimedia technological infrastructure and its associated challenges through the perspectives of the Global Majority communities. It was developed as a year-long collective investigation with Global Majority movement organizations with tech capacity. The Global Majority Wikimedia Technology Priorities document is opened for support signature.

As indicated at the conclusion of these Global Majority Wikimedia Technology Priorities, there is a request to "initiate discussions with the Wikimedia Foundation, affiliates and other Wikimedia Movement bodies to advocate for the incorporation of these priorities into their strategic plans and to have an opportunity to jointly discuss performance indicators for each one of them". Humbly acting on behalf of the collective action that produced this set of priorities, I kindly ask they are considered into this Annual Plan and other WMF tech-based deliberations.

Thank you. -- JPeschanski (WMB) (talk) 04:59, 19 April 2024 (UTC)Reply

Thanks for sharing! The ideas in the report resonate with a number of current and upcoming initiatives at the Wikimedia Foundation, including changes around the community wishlist survey, our global developer support, and co-creating tools alongside communities around the world. Let me know if there are any of these topics you'd like to talk more about, and I'd be happy to connect you to the person working on them. KStineRowe (WMF) (talk) 15:14, 2 May 2024 (UTC)Reply

Major Gifts

edit

I think that the Wikimedia Foundation should not rely on Major Gifts and a culture of Philantropy. Instead it is important to try to get the things done having less money. There is from my point of view potential to be able to do that. One thing what is mentioned in the technology strategy is the shifting of the technology development for the Wikimedia Projects away from the Global North. As I wrote in the years before I think the salaries at the Wikimedia Foundation are high. After I got explanations how they are set I understand it. Still it is from my point of view great if it will be possible to be able to reduce salaries without paying people low salaries who make it difficult to live. Hogü-456 (talk) 19:40, 24 April 2024 (UTC)Reply

Thanks @Hogü-456 for your comment. The annual plan for fiscal year 2023-24 (shared last year) discussed our approach to reducing expenses, as part of a multi-year strategy to slow overall growth in line with our revenue predictions for the coming years. This draft plan for the next fiscal year, provides further details about how this is being done including investments in infrastructure and prioritising support for the movement. Regarding our workforce, the Wikimedia Foundation is globally distributed with staff (49% of the workforce is based outside of the US) and we recruit in over 50 countries. You can read about our approach to compensation for a global workforce and global guidelines on Diff. KStineRowe (WMF) (talk) 15:18, 2 May 2024 (UTC)Reply

Employer of Records Fee

edit

How high is the fee the Wikimedia Foundation needs to pay if a person is working for the Wikimedia Foundation outside of the US and is employed through a employer of record. I have thinked about if it would be cheaper to employ people who work for the Wikimedia Foundation and that are located in countries with a huge chapter like in Germany at the local chapter instead of a employer of record. To get the cost covered the chapter would then have a agreement with the Wikimedia Foundation. Hogü-456 (talk) 18:27, 1 May 2024 (UTC)Reply

Hi @Hogü-456, thanks for your suggestion. From a purely cost perspective, Employer of Record (EoR) agreements benefit from economies of scale. Consolidating employment under one or two EoRs allows for streamlined administration, compliance with labor laws, and access to specialized expertise in managing global and local employment arrangements. Secondly, we intentionally try to balance cost-saving measures with the need to provide an equitable experience for all staff. Working with an EoR allows us to provide a more consistent hiring and benefits experience across many different countries. Relying on local chapters to manage employment could introduce inconsistencies in employment terms, benefits, and protections across different regions, which would raise concerns about fairness and make it harder to recruit and retain top quality staff - pushing up overall costs to the organisation. All in all, using an EoR helps us to have a more diverse and globally distributed workforce, not less. KStineRowe (WMF) (talk) 15:32, 17 May 2024 (UTC)Reply
@KStineRowe (WMF) from my point of view inconsistencies in employment terms, benefits and protections across different regions should be avoided. If you think an equitable experience for all staff is not so propable when the employment happens through local chapters it is form my point of view important to check where there are differeces at the moment. So I mean countries with different conditions for employees of the Wikimedia Foundation and ones of the local chapter located in the same country. Hogü-456 (talk) 19:14, 21 May 2024 (UTC)Reply

Confusing analysis of the data

edit

Hello, maybe it's because I'm not a native English speaker, but I don't understand this sentence very well: "While the ultimate impact of AI remains to be seen, the impact of this pressure over time may contribute to some of our declining metrics like regional traffic and new contributors." And then, there are several statistics being provided. However, the regional metrics shows a decrease only in three regions (Subsaharan Africa, LAC and MENA), but then in all the other regions it has actually increased. So are there declining metrics or not? Then, the "new contributors" link does show a slight decrease in active editors, but it does show a 10% increase in content interactions. The data is showing one thing, however, the sentence is saying exactly the opposite from what the data is showing. Moreover, none of the metrics being provided show a correlation between the perceived threat of AI and the declining metrics. This is a construed correlation, because there's no further information indicating this in the metrics or in the analysis. Is there a more lenghty analysis on what's really going on? Scann (WDU) (talk) 22:37, 8 May 2024 (UTC)Reply

Hi @Scann (WDU), thank you for this question.
We do not have evidence that establishes a correlation between AI and our metrics. What we mean to communicate is that we are monitoring trends relating to AI and have some concerns. One concern we have is that, if people rely on chatbots and other AI products (many of which source their information from Wikipedia), they may not visit Wikipedia. And if fewer people are visiting or aware of Wikipedia, we worry that we will not get as many new contributors over time who will help sustain our projects.
We have seen declines in unique devices in certain regions - most notably Latin America & the Caribbean, where traffic has been declining for some time. We linked to the Q2 snapshot of unique devices, but in our monthly reporting you can see the overall trends more clearly - see page 5 of the Movement Metrics report from February (this is part of the same report we linked to for "new contributors").
As you noted, traffic has increased in some regions compared to the prior year. However, in regions with the most traffic – Northern & Western Europe, North America, and East Southeast Asia & Pacific – the longer term trend has been largely flat or declining traffic. In North America we observed unexpected increases in the beginning of 2024 that we are investigating; one hypothesis is that the increase is due to increased automated or spider traffic that is being miscategorized as user (human) traffic. And one trend that is more clear: automated and spider traffic has been increasing dramatically over the past couple of years. See, for example, the pageviews from automated and spider traffic on English Wikipedia and Spanish Wikipedia.
The "new contributors" link should go directly to page 7 of the Movement Metrics report from February, where we noted that new active editors decreased significantly by about 9.5% compared to the same time last year. We have been observing this decline in new editors for awhile now. We have investigated several hypotheses but so far have not identified a clear answer for this decline. KZimmerman (WMF) (talk) 18:53, 10 May 2024 (UTC)Reply

MENA Community call

edit

This is a section for questions left over from the MENA community call held on May 4, 2024.

1. The editing community in many regions doesn't see an immediate benefit in affiliates, hubs, or other governance structures. Was the editing community considered?

At the Wikimedia Foundation, we consider all parts of the movement and strive to include them in annual and long-term planning conversations, whether they are individual members of the editing community, affiliates or another group within the movement. This year the planning began through Talking:2024 and unsurprisingly we did hear varying perspectives on this.
Some noted how the editing community does not see an immediate benefit in affiliates. Others mentioned the good work by affiliates who are deeply engaged with the community (organising events and trainings, working with partners on content donations, engaging in advocacy locally, providing microgrants for contributors to cover their books, tickets, to write articles etc). Most recognised the complex task at hand - how do we bring together a community that is so huge. How do we ensure everyone feels like they belong and are heard?
The need for individual volunteer representation within the Movement Charter was also brought up at the recent Summit in Germany and voted upon by the affiliates present (Wikimedia Foundation staff and trustees did not vote on these). The  Wikimedia Foundation feedback on the final draft of the Movement Charter also talks about the need to enable the movement to effectively handle present and future challenges (for example, more external regulation, generative AI, and graver risk of external interference in sharing free knowledge). मयूर Mayur Paul (WMF) (talk) 18:20, 21 May 2024 (UTC)Reply

2. From the summary's text: “the Foundation will also begin to signal more clearly where some recommendations are more informative than others in our collective pursuit of the 2030 Strategic Direction.” How will the Foundation begin this? Who will perform this task?

The Foundation is already working on several of the movement strategy recommendations and has been since the recommendations were finalized. You’ll find these recommendations throughout our Annual Plan. For example, Recommendation #2: Improve User Experience informs our Infrastructure goal, especially through investments in the contributor experience, an improved product experimentation platform, MediaWiki sustainability, and shared movement metrics.
Recommendation #4: Ensure Equity in Decision Making informs our Equity goal, including work on movement governance, resource distribution, closing knowledge gaps, and connecting the movement.
Recommendation #3: Provide for Safety and Inclusion informs our Safety & Integrity goal, which involves work around trust & safety, human rights advocacy, and protection against scaled abuse.
Recommendation #10: Evaluate, Iterate, and Adapt, informs our Effectiveness goal, which involves our work on financial sustainability, improving employee engagement and effectiveness, and streamlining Foundation processes.
At the same time, there are other movement strategy recommendations that other groups throughout the movement may be better positioned to lead on, and we invite thoughts on that here onwiki and also on the conversations we're joining with communities around the world. NGunasena (WMF) (talk) 15:26, 22 May 2024 (UTC)Reply

3. Given the shift towards chat-based search and AI-mediated content consumption, how does Wikimedia plan to maintain visibility and attribution for its content in this evolving internet landscape?

You raise an important question and one that Wikimedia Foundation teams have been thinking about a lot. We want internet users to know where the information they encounter is coming from, regardless of whether they're using a search engine, a smart speaker, or a conversational AI. We’ve long asked internet companies and publishers to use Creative Commons’s “T.A.SL.” framework for proper attribution– which asks content re-users to identify the Title, Author, Source, and License for Wikimedia content they re-publish. In November 2022, the Foundation actually extended our brand attribution guidelines for AI platforms as well.
In this next FY, the Legal, Product, Brand, and Wikimedia Enterprise intend to actively share these attribution guidelines with conversational AI platforms. We want to explicitly encourage developers to link back to their Wikimedia content sources whenever possible. Our hope is that getting AI platforms to add proper attribution will be good for Wikimedia, in driving traffic back to the platform and visibility for our brand, and good for the AI platforms, in allowing their users to review and evaluate sources. NGunasena (WMF) (talk) 17:56, 23 May 2024 (UTC)Reply

4. With regards to the second goal, Equity, does content growth only correlate with tools and support systems that are easier to access, adapt, and improve? Or should the editing skills be improved?

Thank you for this question. I fully agree with your premise that both tools and skills are needed to increase volunteer content creation.
As the platform provider for Wikipedia and the sister projects, WMF is uniquely positioned to do certain things well. Improving the tools and the platform that volunteer editors use to create content is one of the things that we can and must do to support this movement. It is where we believe we at WMF can have the most impact. In many cases, Affiliates and online organizers, not WMF, are better positioned to do the skills development that you mention. Affiliates and organizers are often closer to editors who need this kind of support. They know specifically what skills need to be developed, and the specific contexts in which the editors operate (eg., what is needed to support editors in MENA may not be the same as what is needed to support editors in other regions).
There are, however, areas where I believe WMF can play a supportive role for Affiliates and organizers to do this important work. This is where the “support systems” piece comes in. For example, we continue to support Affiliates, organizers, and other experts to teach each other through programs like Let’s Connect. In this case, WMF provides the platform and infrastructure support, but it is Affiliates, organizers, and other community experts who teach other how to uplevel certain skills. You may also have heard of the Africa Growth Pilot where we are working with communities to understand what kind of editing skills training might help to increase editor retention. As we learn from this pilot, which is currently only happening in Africa, we can think about how we might support communities to adapt and contextualize these kinds of training to their local communities. RWeissburg (WMF) (talk) 16:40, 28 May 2024 (UTC)Reply

5. With regards to the fourth goal, Effectiveness, how can the affiliate ability to multi generational budget and planning be ensured?

For Affiliates to be able to support volunteers in their communities, they need to be able to carry out work in the near-term while also planning for the future. We have heard from many Affiliates that while they may be able to fundraise from outside of the Wikimedia ecosystem (see here for how and where Affiliates are raising funds), they need more clarity on what WMF funding will look like over the coming years. To do this, we’ll be working with Regional Funds Committees and Affiliate representatives to share a 3-year view of WMF grants, broken out by region. This will allow Affiliates to understand what the near future looks like in terms of funding and make plans for how to sustain and grow their work beyond these funds if that is a priority for them.
Program Officers from each region will continue to work with Affiliates to engage in longer-term strategic planning, including but not limited to financial planning. RWeissburg (WMF) (talk) 16:40, 28 May 2024 (UTC)Reply


6. With a slowing rate of revenue growth from online fundraising, what innovative approaches is Wikimedia exploring to ensure long-term financial sustainability beyond traditional donation models?

Thanks for asking. The Revenue Strategy section (Arabic translation here) of the annual plan offers a lot of detail on this question. In short, we are looking to diversify our revenue sources in areas that are more resilient to changing readership trends, such as through the Wikimedia Endowment and Wikimedia Enterprise, as well as growing our major gifts program to continue building resilience for the long term. MeganHernandez (WMF) (talk) 08:39, 24 May 2024 (UTC)Reply

7. The Arabic region faces a lot of challenges in terms of safety and conflicts. How does Wikimedia plan to measure the impact of its projects and initiatives in regions affected by conflict?

The Foundation is working to expand the community safety surveys, which will help both communities themselves and the Foundation to better understand the impact real world events might have on the projects alongside other factors. The organisation is also continuously working to improve measurements of the community projects impact on the world they serve but currently does not aim to build infrastructure specifically related to conflicts as the latter could create considerable risks for communities in affected areas due to the transparent platform environment where data is available (stats.wikimedia.org, etc.). --Jan (WMF) (talk) 13:13, 27 May 2024 (UTC)Reply

8. “These changes are accelerating as the internet has become more diverse and global, and less English-dominated.” This domination could be due to the immediate and machine translation of pages. Does Wikimedia plan to make these translations available permanently and for all of its pages?

Thank you for the question. The changes referenced in annual plan text refer to a global shift away from link-based search architecture and toward a chat-based search architecture, which is still in its early days. These changes involve how people find information across the internet, and are driven by a number of global trends including the availability of new artificial intelligence technologies and the rise of short form, personality driven content funnelled to consumers by complex customization algorithms. They impact the Wikimedia projects, but reach much more broadly to all internet based platforms and content. For the Wikimedia Foundation’s Machine Translation service (MinT), The Foundation is working on diversifying its use to ensure more sustainable support for languages and knowledge in our projects through our MinT for Wikipedia exploration work.
As part of an initial exploration MinT for Wikipedia will support readers in different Wikipedias by giving them the option of easily reading machine-translated articles that are available in other languages in their choice language.
This can be seen in a special page in any Wikipedia ( e.g., English Wikipedia). The entry points are not yet available, and those will be enabled gradually through an iteration cycle as the feature becomes ready for broader exposure to users.
Furthermore, we will explore integrating it into other projects based on what we learned in through its release on Wikipedia. We encourage your region to follow this work.--Runa Bhattacharjee (WMF) (talk) 06:10, 29 May 2024 (UTC)Reply

9. Artificial Intelligence is a major concern, not only for WMF and the movement as a whole, but also for all companies and institutions, and the development is very rapid. I know the organization is working and experimenting with ways to adapt AI to support the mission, and not be an obstacle to achieving it; what are the plans of the Wikimedia Foundation towards achieving such a huge need?

As a Foundation, we’ve had a dedicated machine learning team since 2017 and have been working with volunteers on machine learning tools to help support the work of editors. We’ve already introduced machine learning tools that support editors and help support our mission, such as models for measuring revert risk for different edits or the Content Translation tool, which integrates with external machine translation models and has been used by volunteers to help translate more than 1.5 million articles.
Our approach to AI and machine learning is based on centering the work of the people behind the Wikimedia projects. Our work on AI starts from this shared value, and shapes both our own use and tooling with AI as well as our advocacy and research work in this field. There are a number of ways that the Foundation is thinking about and using AI and machine learning as part of our work. Under our current year’s annual plan, we’ve focused on modernizing our machine learning infrastructure to support mission-aligned ML tool use on our projects. This has included experimenting with whether and how we can serve reliable, verifiable knowledge via off-platform AI assistants like ChatGPT. We have also experimented with how machine learning might be used to help smaller wiki communities automatically moderate incoming edits.
For the year ahead, we’re looking at using machine learning in service of volunteer’s efforts to identify and address information, as well as helping policymakers and policy influencers seeking to regulate the use of artificial intelligence understand how AI and machine learning tools are used to support volunteers in improving encyclopedic content quality and fighting disinformation. We are also looking at opportunities to continue advocating for attribution and credit for the work of Wikipedia volunteers in large language models. NGunasena (WMF) (talk) 05:36, 30 May 2024 (UTC)Reply

10. What strategies is Wikimedia employing to engage a new generation of readers and contributors, particularly those who may prefer different forms of interaction with knowledge, such as generative AI tools like ChatGPT?

Thanks for the question! Generally, we are looking at a number of different possible areas of opportunity for engaging new readers and contributors. Some but not all of these include generative AI.
In the generative AI sphere, we’ve experimented with a plugin for ChatGPT that uses Wikipedia as its main source. Rather than a product idea, we used this as a way to learn more about younger audiences and the way they think about connecting AI and Wikipedia. Currently, we’re working with a professor from the University of Mississippi who conducted an experiment using this plugin in a classroom setting with undergraduate students to write a paper. However, neither of these experiments have so far provided compelling evidence to indicate that younger audiences are shifting towards generative AI for the type of information typically found on Wikipedia. That said, we plan on digging deeper into this in the future. Similarly, we’re currently running experiments on Video question and answer (Q&A) which focuses on exploring different media formats for delivering information for users, as well as an experiment that allows users to add facts to Wikipedia using AI, aimed at making contributions easier for new generations.
Within the website and apps, we are planning on experimenting with different strategies around making the site easier to browse through recommendations and new browsing and content discovery experiences. We believe this will make it more likely for new generations of readers to return to Wikipedia frequently when they wish to learn or are interested in more open-ended learning. We are also interested in making it easier for readers to find the content they are interested in, or in surfacing relevant content quicker. To begin this work, we plan on experimenting with using models that summarize or simplify the content to make skimming and getting quick context easier across articles.
If you have any further questions, feel free to reach out to us here. OVasileva (WMF) (talk) 07:14, 10 June 2024 (UTC)Reply

11. How is Wikimedia strengthening its defences against disinformation and ensuring the integrity of its projects amidst increasing content veracity challenges and regulatory pressures?

Great question, and a super important topic. We’re dedicating substantial resources to this work next year, which is described in our “Trust and Safety” objective, which is part of our overarching Wiki Experiences (WE) work about how users engage with the Wikimedia platforms. There are three main workflows (listed as “Key Results” or KRs) under this objective.
  1. WE4.1 is about finding ways to measure harassment and harmful content on our platforms, and then identifying some countermeasures that are also in accordance with evolving global regulations.
  2. WE 4.2 will develop new approaches to detect bad actors and sockpuppet accounts on our platform while also decreasing the collateral damage to good faith users.
  3. WE 4.3 involves working to improve infrastructure, tools, and processes that protect Wikimedia against scaled and direct abuse, which we’ve seen on the rise in recent years. ACooper-WMF (talk) 20:08, 11 June 2024 (UTC)Reply

Artificial intelligence claim

edit
By many estimates, English Wikipedia forms the largest single source of training data ingested by ChatGPT – and among the highest-weighted for quality.

Where does the cited WaPo source make this claim? Was it intentional to link to the same article twice or was a different source intended here? czar 02:38, 23 May 2024 (UTC)Reply

This NYT article makes some closer claims about importance but still not as wide-reaching as being ChatGPT's largest data source (isn't that Google Patents per token count and the WaPo article?) and highest-weighted for quality (not sure what that would mean from the source). czar 04:19, 23 May 2024 (UTC)Reply

@czar, thanks for the question. The Washington Post names us as one of the top datasets for training large language models. The other top datasets it lists also include Wikipedia content within their data sets (Common Crawl and C4). (CC has 20% URLs from Wikipedia, C4 has 19% of tokens from Wikipedia.) This paper shows that en.wikipedia.org and en.m.wikipedia.org are among the top-3 domains in the corpus in terms of number of tokens, and goes into further detail into how much we make up of C4’s dataset. The Foundation’s research team also wrote a paper where they discuss that in more detail https://arxiv.org/abs/2204.02483 NGunasena (WMF) (talk) 05:38, 30 May 2024 (UTC)Reply
@NGunasena (WMF), thank you! "One of the top datasets for training large language models" (like ChatGPT), as you've written, strikes me as a much better paraphrase of WaPo's claims than what's currently in the plan, since the WaPo source doesn't say that Wikipedia is either ChatGPT's largest single source or the highest weighted for quality, as the plan claims the source says. czar 11:53, 30 May 2024 (UTC)Reply
I've gone ahead and made that change. Thanks @CzarCZAR. CAlbon (WMF) (talk) 23:18, 10 June 2024 (UTC)Reply

Metrics for top goals

edit

Thoughts on two of the proposed metrics:

[A] "Increase # of quality articles on Wikipedia" -- This should not be a target metric, though progress in related goals would hopefully show up in the quality of information readers and reusers get from the projects.

  • "Quality" isn't well defined (even en:WP:QA is now only historical) has many meanings, thanks for linking to the one being tracked globally!
  • On larger projects, average quality increases organically thanks to community work, so this metric isn't diagnostic.
  • Quality in and of itself isn't our core strength, compared to coverage, speed + timeliness, participation, efficient curation.

To the exclusive focus on Wikipedia:

  • Increasing the # of articles across some threshold is not the top need, though it a healthy metric for the project.
  • Wikipedia is not necessarily the part of our knowledge ecosystem that needs the most work.
  • Wikipedia depth is not the area where we can make the most impactful change (in coverage, readership, and reuse) -- though as an umbrella do get updated to reflect improvements in other areas (data, media, &c).
  • Overindexing on "quality" leans into the potentially harmful trope that Wikipedia's mission is to be the world's reliable source of knowledge, whereas absolute reliability is an intentional structural weakness, accepted as a tradeoff with other strengths.

Potential metrics that could more effectively get at the core issues:

  • Coverage of new advances / top requests; balanced breadth across essential areas; tools to visualize and fill gaps (which would define helpful metrics); existing cleanup categories (already-defined metrics, many with their own scripts and tools that can be improved); better + faster integration of work on a given topic across Projects. This is closer to existing editorial goals, amplifies our strengths, doesn't overprivilege WP.
  • The possibilities for automation these days are spectacular; future Wikipedia (or its successors) will offer contributors much higher returns on invested time. That seems in the critical path. Speed/leverage of people working on one of the above metrics, and sufficiency of capacity to the size of a backlog, may be a useful second-order metric.

[B] "Maintain our programmatic expense ratio at 77%" -- this is tame for a top-level goal.

  • Shouldn't this be increasing? We just saw a period of rapid growth, and can now consolidate.
  • 23% overhead isn't great, considering that much of social + governance overhead is classified as programmatic expenses.
  • Some programmatic expenses (process creep, projects running hot and cold, long consultations for small effect) are a net negative for the community, taking time away from project work for little benefit. Finding ways to make that more efficient and satisfying, and to gently increase the expense ratio, seem like goals within reach.

SJ talk  15:55, 23 May 2024 (UTC)Reply

I'm not sure what "We will increase the number of quality articles on Wikipedia" means, exactly. Does it mean articles that have been chosen by the community in the various language versions of Wikipedia to formally receive a status similar to English Wikipedia's "good article" or German Wikipedia's "Lesenswerter Artikel" (the status one level below "featured article")? Or is the "quality" measured by some other means; if yes, what means could that be? Gestumblindi (talk) 09:47, 24 May 2024 (UTC)Reply
Hi @Sj and @Gestumblindi, I revised the metrics section to add a link to our working definition for quality articles (thank you for flagging this!). This working definition for quality comes from our work on the Knowledge Gaps Index (specifically the language-agnostic quality model). The model was designed to identify actionable levers for improving quality (such as references, images, and intra-wiki links).
@Sj, I also made a revision to clarify how we will measure success. Although our overall goal is to increase the number of quality articles on Wikipedia, we will measure our success based on achieving the key results outlined under the Equity goal. (For more details, see Wiki Experiences objective 2 and its key results: 1, 2, 3, and 4.). I'm working on responses to some of your other points; more to come. KZimmerman (WMF) (talk) 22:08, 28 May 2024 (UTC)Reply
Thanks so much KZimmerman! I'd forgotten about that quality measure, and that whole gaps + metrics roadmap. The visualizations of the gaps are also appreciated, please let the communities know how we can help get visuals for the other measures. The key results under objective 2 all seem attainable and important, without the potential externality of a push for e.g. higher thresholds of quality at the expense of breadth. –SJ talk  22:39, 28 May 2024 (UTC)Reply
@Sj When you say "visuals for the other measures," do you mean the annual plan metrics, or some of the measures you suggested? I'd love to collaborate on better communicating & visualizing key needs and areas of opportunity, especially where there's particular interest from community members! KZimmerman (WMF) (talk) 23:32, 30 May 2024 (UTC)Reply
Hi @Sj, thank you for your questions and suggestions! As promised I’m circling back with more responses.
Regarding [A]:
We think quality – and verifiability – are necessary for ensuring that our content has a positive impact on people’s lives. Tracking changes in quality content is one way we can measure whether we are effective in addressing disinformation, misinformation, and missing information in our ecosystem.
Our focus is primarily - but not exclusively - on Wikipedia and the MediaWiki infrastructure that supports all projects (we have identified a critical need to invest in MediaWiki so that it can provide the creation, moderation, storage, discovery, and consumption of open, multilingual content at scale).
As I noted in my earlier response, we have a working definition for Wikipedia article quality that comes from our work on the Knowledge Gaps Index (specifically the language-agnostic quality model). We are using the features in the model to inform the development of tools and processes that will support the growth of quality encyclopedic content.
There are plans to support some of the ideas you mentioned. For example:
  • “Coverage of new advances / top requests; balanced breadth across essential areas; tools to visualize and fill gaps” – some of the planned initiatives under Wiki Experiences objective 2 and its key results (1, 2, 3, and 4) specifically target this. And there’s the work done under the Knowledge Gaps Index to help visualize the gaps.
  • “The possibilities for automation these days are spectacular; future Wikipedia (or its successors) will offer contributors much higher returns on invested time. That seems in the critical path. Speed/leverage of people working on one of the above metrics, and sufficiency of capacity to the size of a backlog, may be a useful second-order metric.” – The work on Automoderator, started this fiscal year, is one step in that direction. And the work under Wiki Experiences objective 1, key result 3 will also support contributors’ capacity.
I’ll share your suggestions with product managers and the researchers and analysts supporting their work - we are always looking for better ways to measure our impact! I’m curious to learn more about your idea around “better + faster integration of work on a given topic across projects” - can you give me an example of the kind of thing you’re thinking about?
Regarding [B]:
The programmatic expense ratio is one measure where we have external benchmarks to assess our performance. Charity rating organizations use this as a measure of effectiveness as well. For example, Charity Navigator gives full points when scoring charities if their ratio is above >70%. Our target is significantly above that and is part of the reason we have a 99% rating with Charity Navigator.
We noted under our Revenue Strategy, “As the Internet has evolved and people are increasingly consuming Wikipedia content off our platform, it became clear that our revenue strategy could no longer primarily depend on donations made on Wikipedia. We began developing new revenue channels that were more resilient to changing readership trends: Wikimedia Enterprise, the Wikimedia Endowment, and a growing Major Gifts program.”
Running fundraising banners on Wikipedia has been an efficient way to invite our wide reader base to donate, but we must also invest in further developing other revenue streams. And although we deliberately increased our ratio of fundraising expenses this year, we were able to realize some efficiencies to reduce our General & Administrative expenses. KZimmerman (WMF) (talk) 23:23, 30 May 2024 (UTC)Reply

South Asia Open Community call

edit

1. What is the plan for the MinT translation model? Are we planning to sunset the project?

Thank you for the question. MinT provides extensive translation support in the content translation tool in more than 40,200 language pairs using different models. It also provides translation suggestions for communities that translate our software and multilingual pages in translatewiki.net, Wikimedia Meta-wiki, MediaWiki.org and more.
There are no plans to sunset MinT. Instead, the Foundation is working on diversifying its use to provide more sustainable support for languages and knowledge in our projects through the MinT for Wikipedia exploration work. The exploration work is in its initial stage. It will support readers in different Wikipedias by allowing them to easily read machine-translation articles available in other languages in their chosen language.
You can see the exploration work in a special page (Special:MinT) in any Wikipedia (e.g. on English Wikipedia). No entry points are available yet, and the feature will be enabled gradually through an iteration cycle as it becomes ready for broader exposure to users.
We also plan to integrate it into other projects based on what we learned through its release on Wikipedia. You can follow this work to learn more about our plans for MinT.--Runa Bhattacharjee (WMF) (talk) 06:26, 30 May 2024 (UTC)Reply

2. Is the Foundation planning to add any plans toward protecting misuse of Commons images without citation?

There aren’t any new initiatives on this front planned for the next fiscal year, but our legal enforcement team can sometimes help with individual cases. Since the copyright in Commons images is retained by the original author, the Foundation doesn’t have as many enforcement tools as with other legal issues (such as trademark infringement). If the misuse is particularly prominent or widespread, though, there may be options available to us. You can report potential misuse to legal-tm-vio@wikimedia.org. On the technical side, we have made improvements to the UploadWizard that allows users to be more aware of content copyright while uploading, which can help in better visibility of the rights associated with the content. --Charles M. Roslof (WMF) (talk) 18:48, 31 May 2024 (UTC)Reply

3. How is the Foundation planning to provide legal aid to volunteers?

Hi all, I reviewed this question and can provide some information. As a part of our regular legal work, the Foundation can help Wikimedia users involved in legal actions related to their volunteer work with Wikimedia projects with financial support or assistance in finding a legal contact in certain instances. The Foundation has three programs for this work, which are available as standing programs evaluated based on individual requests. The Wikimedia Foundation Legal Fees Assistance Program is for users who are the subject of a legal action because of their elevated or functionary role on the Wikimedia projects such as administrators, checkusers, or similar. The Defense of Contributors program is for good-faith editors who are the subject of legal action because of their editing work. The Community Health Legal Defense program is for community members subject to severe harassment because of their work with the Wikimedia movement.
If you have a case that would fit under one of these programs, you can reach out to the legal team through legal @ Wikimedia.org to contact us. If you are interested in one of these programs, please provide as much detail as possible: if you are the subject of a legal action, we typically need a copy of any court documents you received to evaluate the case for eligibility under these programs. -Jrogers (WMF) (talk) 17:56, 31 May 2024 (UTC)Reply


4. Can the Wikimedia Foundation more regularly share information about new and updated tools for language and content translation with Indic communities? There is a knowledge gap among the content community about what tools exist and how to use them.

Hi, I'm Uzoma, I support community outreach and communication for the Language Team at the Wikimedia Foundation. We greatly appreciate your interest in staying updated on new and updated language tools and your contribution to our projects.
The WMF Language team communicates language tool-related updates and works with the movement through newsletters, blog posts, project documentation pages, virtual meetings, community village pumps and Telegram channels.
The Indic community is welcome to follow our work and get up-to-date information by:
As mentioned above, the Language team also communicates specific changes and enablement plans in the Wiki's village pump and to community members' user talk pages and emails when necessary.
Do these communication methods meet your needs? Are there other or better ways we can connect with Indic language communities? Please let us know your preferred communication method if it differs from our existing method and channel. Thank you! UOzurumba (WMF) (talk) 16:31, 31 May 2024 (UTC)Reply

5. What are the goals and metrics planned for improving the editing experience for editors?

Great question. There are two broad objectives/goals – and metrics associated with each – designed to improve the editing experience for editors, across experience levels.
I've detailed them below. If anything you see brings new questions to mind, please let us know.
Editor-focused goals for 2024 - 2025
  1. Contributor experience: Both experienced and new contributors rally together online to build a trustworthy encyclopedia, with more ease and less frustration.
  2. Encyclopedia content: communities are supported to effectively close knowledge gaps through tools and support systems that are easier to access, adapt, and improve, ensuring increased growth in trustworthy encyclopedic content.
Editor-focused initiatives and metrics for 2024 - 2025
As part of the "Contributor experience" goal, the Product and Technology Department will invest in the following:
  • Connection among editors: invest in the ways that contributors can find and work together with other contributors who are interested in the same things, with work that may include tools for WikiProjects and in on-wiki communication.
    We will consider this work effective if we introduce or improve ≥1 workflow that helps contributors with common interests connect with each other and contribute together.
  • Workflows for newcomers: continued development of editing experiences that provide guardrails, suggestions, and learning for newcomers so that they can be constructive quickly.
    To evaluate the impact of this work we will monitor changes in the percentage of newcomers who publish ≥1 non-reverted edit in the main namespace on a mobile device, as measured by controlled experiments.
    This metric is also referred to as constructive activation, the baseline for which we are establishing in T360829.
  • Dependability: streamlining and repairing the tools that allow patrollers and editors with extended rights to do their work without frustration.
    Through this work, we are striving to increase editors' satisfaction with ≥4 moderatortion experiences by __%.
    Note: exact percentage increase in editor satisfaction is still being defined.
As part of the Encyclopedia content goal, the Product and Technology Department will invest in the following:
  1. Increasing coverage of quality content in key topic areas
    To evaluate the impact of this work, we measure the growth in new articles in key topic areas (still be defined).
  2. Improving the technical infrastructure to onboard new languages
    We will consider this work effective if by the end of Q2 we have implemented at least two recommendations, both social and technical, to support languages onboarding for small language communities, with an evaluation to analyze community feedback.
  3. Growing access to quality source material to close strategic content gaps
    We will consider this work effective if by the end of Q2 we introduce 2 new features that guide contributors to add source materials that comply with project guidelines, and 3-5 partners have contributed source material that addresses language and geography gaps.
  4. Improving workflows to scale quality content across the movement, especially in smaller language communities.
    We will consider this work effective if by the end of Q2, we've enabled Wikifunctions calls on at least one smaller language Wikipedia.PPelberg (WMF) (talk) 22:32, 3 June 2024 (UTC)Reply

6. There is no uniform rule across wikis for defining a stub, and practices can vary widely. How can this be made more uniform?

The content policies, guidelines, and norms of each wiki are locally defined by their volunteer communities, and are not centrally managed or controlled by the Wikimedia Foundation. Volunteers can use shared spaces across Wikimedia projects, such as Meta-Wiki or off-wiki platforms and events, to discuss and learn new practices from other wikis that they would like to implement on their home wiki and to improve things like defining a stub. KStineRowe (WMF) (talk) 19:50, 3 June 2024 (UTC)Reply

7. There is systematic bias on English Wikipedia where efforts/edits are deleted. The notability criteria are questionable and more equity needs to be brought in. Similarly, many Indian sources don’t meet requirements on English Wikipedia. What can WMF do to ensure greater equity in these criteria?

The Wikimedia Foundation does not play a role in content decisions, though it does build software that helps users understand and use notability criteria in order to reduce the likelihood that their edits are reverted. You can learn more about this work on the Edit Check project page.
Additionally, Notability criteria determine which topics have gained sufficiently significant attention by the world at large over a period of time in order to qualify to have their own standalone articles. This criteria is defined by the editing community of each Wikipedia project, not the Wikimedia Foundation.
For reference: In English Wikipedia, a topic is presumed to be notable, and therefore suitable for a stand alone article, when it has received significant coverage in reliable sources that are independent of the subject. There is a page for both the General Notability Guidelines, and guidelines for specific subject areas. Similarly to the notability guidelines, the rules around reliable sourcing are set and enforced by volunteers. On English Wikipedia in particular, there have been concerns around paid reporting in some Indian news outlets, which are known to create challenges for editing on topics primarily covered by those sources. The Perennial Sources list offers some clarity on which sources are considered credible and reliable, and how volunteers can help improve the list. Editors can also discuss source reliability on an article’s talk page. KStineRowe (WMF) (talk) 19:54, 3 June 2024 (UTC)Reply

8. What specific parts of the annual plan are relevant to the South Asian movement, and how can we engage in the annual plan better?

Over the last few years we have continued to evolve and improve a collaborative planning process. The four overarching goals (Equity, Infrastructure, Safety, Effectiveness) remain constant, while the work and deliverables within them iterates on the significant progress made in the current year.
This year the plan began with Talking:2024 - conversations between Foundation staff, leaders, board members, and Wikimedians around the world including in South Asia. It looks at external trends and what the world needs from us.
As South Asia has a thriving community all of the Goals of the Annual Plan are intrinsically relevant. Depending on interests, some highlights and examples South Asian contributors may want to specifically look at are below:
E.g. Our work on product and technology within the infrastructure goal
  1. How we can support a new generation of volunteers and increase the number and availability of smaller, structured, and more task-specific editing workflows (E.g. Edit Check and Structured Tasks).
  2. What we are doing to support languages onboarding for small language communities, with an evaluation to analyze community feedback.
  3. Creating new features to guide contributors to add source materials that comply with project guidelines, and have 3-5 partners contribute source material that addresses language and geography gaps.
  4. How we are building a new Wishlist that better connects movement ideas and requests to Foundation product & tech activities.
E.g. Our work in the Equity Goal
  1. Looks at Movement Governance and Decision-Making - specifically with a focus on shared accountability, equitable participation, and supporting governance success across and for the movement.
  2. Within Resource Distribution - we commit to collaboratively determine the regional allocations of the grantmaking portfolio. We will also work together to take concrete steps towards subsidiarity in resource distribution by moving decision-making about global campaigns and rapid funds grants to those closest to the work.
  3. As part of connecting the movement, we commit to deepening our regional connections through multicultural and multilingual communications to build two-way conversations informed by local knowledge. Our regional specialists will build on and maintain the many personal relationships they have forged in order to cultivate collaboration and shared understanding with local communities. We will listen and amplify local regional ambitions, stories, gatherings and more - beginning with this annual plan itself where we will use the process of engaging with the plan to elevate regional priorities and plans this includes the South Asia Open Community Call and more.
E.g. Within our Safety & Integrity Goal
  1. We will protect our People by strengthening the policies and systems that provide for the safety of volunteers. We will do this by strengthening trust and safety on Wikimedia projects, protecting human rights when our projects are at risk, and building technology to prevent scaled abuse. You can see more about legal support in a different answer within this section.
  2. We will protect our projects by defending the integrity of our projects from attempts to limit access to knowledge. We will do this through legal defense and compliance, and countering disinformation.
  3. We will promote the value of Wikimedia's free knowledge model so that more people will help us protect it. We will do this by advancing a positive vision for internet policy that supports and protects the broader digital commons upon which Wikimedia depends.
Hope this helps and thank you for your question.RASharma (WMF) (talk) 19:48, 5 June 2024 (UTC)Reply

Feedback towards the Annual Plan 2024-2025 from Centre for Internet and Society - Access to Knowledge

edit

Hello! Based on the Indic communities' feedback and some personal experiences, we are providing the following feedback towards the Wikimedia Foundation’s Annual Plan 2024-25 under two sub-sections

Contributor experience

edit
  1. What are some of the changes in the existing mobile application and new releases that the Wikimedia Foundation is planning to do in the coming year? What can the community expect from the upcoming releases?
  2. How will the “suggested edits” feature look in the coming year? What changes and improvements can be expected compared with last year?
  3. What changes will be made to the MediaWiki platform and the interfaces to meet Wikipedia's core needs? Are there any plans to degenerate some of the existing features of the MediaWiki platform?

Content and Projects

edit
  1. Are there any plans to differentiate AI-generated images/content from regular images/content? Like adding a tag/label to the images/content?
  2. What are some of the improvements expected to the translation model in the coming year? Will the services like MinT which has one of the models developed by AI4Bharat take a live feed from the user for improvements as feedback for future translations? Or Is this already in place?
  3. Are there any projects where the Foundation is trying to involve AI in bridging content generation where technical communities can contribute to it?
  4. We realise the API has certain limits as it is a strategic step taken by Wikimedia Foundation from the Movement Strategy Recommendations. But is there any chance or space for permission to increase the limits to access it for the developer community in building/developing community dashboards or tools where querying data might not be possible due to the current limitations?

Thank you!

On behalf of CIS-A2K — Nivas (CIS-A2K) (talk) 05:30, 3 June 2024 (UTC)Reply

Hello Nivas and thanks for all these great questions! The topics you raise span the work of multiple teams at the Wikimedia Foundation. I'll reach out and let them know to respond; most likely you'll hear from different staff on different questions as they're able to reply. KStineRowe (WMF) (talk) 21:37, 3 June 2024 (UTC)Reply
Hi @Nivas (CIS-A2K)
As it relates to your question about mobile apps, there will certainly be new releases in the coming year. Our current plans include enhancements to search within the app for improved article discovery, text to speech, exploring topic based reading, experimenting with a Wikiwrapped concept, learning if an alt-text adding suggested edit is feasible, improvements to recognizing our donors in the app and if time permits allowing access to sandbox editing.
Of these the first ideas we are working on search and article discovery enhancements and learning if an alt-text adding suggested edit is feasible.
Certainly open if there are other ideas top of mind for us to consider or if the ones I mentioned above sound particularly of interest to you. JTanner (WMF) (talk) 23:38, 3 June 2024 (UTC)Reply
Hello @Nivas (CIS-A2K)
I'm happy to help answer your questions about Suggested edits.
On the web and mobile web, the Growth team will focus on Structured tasks. Structured tasks are Suggested edits that break down editing workflows into a series of steps that newcomers can accomplish easily. In multivariate tests, Structured tasks have repeatedly been shown to increase newcomer participation (1, 2). We hope to find ways to surface Structured tasks to brand new account holders, while also investigating new types of Structured tasks. You can read more about the Growth team's annual plans here.
On the Mobile app side, the iOS app team is going to explore an alt-text Suggested edit on select wikis. You can read more about those plans here: iOS_Suggested_edits_project/Alt_Text_Experiment.
Are there particular Structured tasks that your community might find helpful? Are there certain improvements to Suggested edits that you are hoping for? We are always eager for feedback, so please feel free to chime in here, or on one of the related project pages. Thanks! KStoller-WMF (talk) 04:11, 4 June 2024 (UTC)Reply
Hello Nivas,
With regards to question (4) - I am assuming you are referring to the “Enterprise API” so I will answer with regards to that. I am making this assumption because that API has been built over the last three years a way that is consistent with two points in the Movement Strategy Recommendations - which is described in detail on the project’s FAQ here. but if I have misunderstood, please let me know.
Are you familiar with the various methods of access to that API service which are detailed on Meta, here? Notably this includes API’s website itself, WikimediaCloudServices access and, the final option which states, “Those who have a non-commercial and mission-relevant use-case, which cannot be fulfilled by existing free-access APIs/dumps etc, can request ongoing access to the paid service at no cost.”
I know this last option is not a “structural” solution to any limitations you might be referring to, but it is as an attempt to proactively respond to unique circumstances which aren’t covered elsewhere. More generally, over the next year or so, the API’s general usage terms is going to be restructured to ensure that the no-cost usage version is not simply a "trial" version, but something that can be used on an ongoing and consistent basis for community/volunteer/research needs. The intention of the paid-use of that API service is for the extremely large-volume commercial users of the WMF's infrastructure to no longer be a technical and financial burden upon the WMF's publicly available infrastructure. But movement/volunteer/mission-relevant use should always be freely available and not negatively impacted. Please contact me on the project talkpage or via its Phabricator board to followup with specifics. LWyatt (WMF) (talk) 21:07, 4 June 2024 (UTC)Reply
Hello @Nivas (CIS-A2K), regarding question 2 about MinT - the researchers or institutions behind these models often release updated models. For example, in the case of IndicTrans2 model, AI4Bharat had released updated models and we integrated these new versions. The updates could be to support new languages, better models trained on bigger/better data, algorithmic improvements or inference optimizations. We actively monitor these developments in upstream projects. We also try to add new models if we find models with compatible license.
Regarding the feedback about improvements, due to the nature of the models (or ML architecture) specific issues like fixing a single word translation is not possible. Generic feedbacks such as abbreviation support, number support, sentence segmentation issues, punctuation issues can be addressed by training data fixups. We use the upstream issue trackers to report such issues. To reduce the user facing issues with say, wrong punctuation, we have a post-processing layer in place in MinT, that is fully in control of the WMF Language team. Runa Bhattacharjee (WMF) (talk) 05:38, 6 June 2024 (UTC)Reply

Igbo Community call

edit

1. What are the actual plans to balance inequality in the grants process? Organisations in the Global South are bombarded with questions and other organisations in the Global North are just asked simple questions and their grants are approved.

In the past three years, we have observed and supported a growing interest in grant requests from the global south across different regions. Our commitment to a participatory grantmaking approach means that community representatives, through regional funding committee structures, review requests and make funding decisions.
We continuously strive to enhance the review process, including the proposal frameworks, to improve overall decision-making. Over the past year, we have noticed that regional funding committees place a strong emphasis on assessing proposals based on their potential for impact and the capacity of the applicants, both at the organizational level and in terms of their knowledge of Wikimedia projects. It is also crucial for the projects to demonstrate a clear alignment between the requested resources and the proposed impact.
To support the committee and staff in gaining a deeper understanding of these capacities, we provide questions and feedback to applicants. The type and quantity of questions may vary depending on the establishment, development, or growth stage of the applicants. This approach ensures that we have a comprehensive understanding of the entity, enabling us to make well-informed and supportive funding decisions.
Our goal is always to facilitate a fair and thorough evaluation process, ensuring that every applicant, regardless of their region, receives the support they need to succeed.--VThamaini (WMF) (talk) 15:36, 11 June 2024 (UTC)Reply

2. Categorised IP blocks are still a challenge for us. People get blocked just trying to edit. If, after all the hard work, they eventually get blocked, the newcomers don't come back. It then becomes online harassment. They should have an action team to ask people what they are doing through pop ups etc to determine good faith before they block people. What is the WMF doing about the categorized IP addresses that place restrictions on some regions (especially the Global South) when they try to edit, causing "editing restrictions" for editors and newcomers?

Hello, my name is Kosta Harlan, I'm a staff software engineer on the Trust & Safety Product team. Thank you for this comment. We agree that IP blocks result in collateral damage that cause problems for newcomers, and that we need to provide other tools and abuse mitigation strategies. IP-blocking-impacts on Phabricator tracks some relevant tasks we are monitoring and working on. In the coming year, as part of WE4.2 in the annual plan, we are working towards improving abuse detection and mitigation strategies so that we have options to more precisely target abuse. We also intend to do more work to establish baselines for monitoring blocks (example: provide structured block reasons) and edit attempts denied due to IP blocks (example: analysis of how IP blocks impact events). If you have other suggestions or questions, please let us know. Thank you! KHarlan (WMF) (talk) 07:43, 12 June 2024 (UTC)Reply

3. On knowledge equity, we have fewer articles about Africans; are there plans to push articles about women into the limelight? Is there also a way to showcase articles from the global south in the limelight to drive page views and awareness?

Hello! My name is Masana Mulaudzi, and I work as the Senior Manager for Campaigns Programs (Organizing) at the Wikimedia Foundation. This is a great question! This year, one of our priorities under the Equity Goal is to support communities “to effectively close knowledge gaps through tools and support systems that are easier to access, adapt, and improve, ensuring increased growth in trustworthy encyclopedic content”. In practice, this goal will help make knowledge gaps more visible to editors, organizers, and partners — with an initial focus on gender and geographic representation — in order to help them to more effectively address those gaps. We envision that by helping our communities close knowledge gaps, readers and users of our projects will be better able to access the information they need. Additionally, we will work on building connections between contributors, which will increase the visibility and discoverability of campaigns, events, and wikiprojects focused on closing specific gaps, including in gender. By providing more nimble pathways for contributors to work together on identifying and closing knowledge gaps in gender and geography, we hope to make it easier for our projects to reflect the contribution of, for instance, notable African women to society. Our hope is that these tactical efforts, combined with our goal to amplify the work of the community in the gender gap space, will collectively add to the profile of the communities that make articles about women possible on our projects.
I would like to add and remind people that WMF, in it's basis, can only STIMULATE and ENCOURAGE. Actual fixing of this problem is something that is the responsibility of and has to be done by, the (future) collective of people writing articles. So relatively speaking, WMFs impact (in numbers) will always be very limited on a problem like this. It can push the scale of society, but how succesful that push is, is very dificult to measure compared to overall societal changes. I think we sometimes forget about that and look a bit too much at WMF for this. —TheDJ (talkcontribs) 13:07, 19 June 2024 (UTC)Reply

Priorities from the Japanese Wikipedia community

edit

In announcing the Annual Planning on wikis, WMF has also posted specific questions with the aim to better listen to the voices of the individual contributors. The post was localized in various wikis and an anonymous survey was added as an additional option to share feedback on the Japanese Wikipedia. Throughout the survey responses a key theme that emerged was around newcomers - ranging from how to better support quality edits to how to bring in new generations of contributors and admins. You can read more on this Diff blog. JNakayama-WMF (talk) 01:15, 21 June 2024 (UTC)Reply

Return to "Wikimedia Foundation Annual Plan/2024-2025" page.