Talk:Product and Technology Advisory Council

Congratulations and thanks

I just wanted to kick off the talk page with a thank you message for setting this up, and for the fine set of folks who have volunteered to serve. - Fuzheado (talk) 13:24, 23 October 2024 (UTC)

Thank you for encouraging this new initiative Andrew ! —TheDJ (talk • contribs) 11:54, 30 November 2024 (UTC)

Any updates?

Now that the WMF Board elections are complete, are there updates for the membership of this committee? I note that there has been no update on the additional affiliate member, which was theoretically going to be announced a while ago. Thank you to all of you who have agreed to take on this task; I think PTAC has a strong chance of being successful with so much knowledge and experience in the room.

Based on the information in the original post, PTAC should have already met once or twice, with an in-person meeting scheduled...any time now? Will minutes or a summary of the meetings be published? I am certain I am not the only person interested in hearing more about the work of the council. Risker (talk) 04:27, 8 January 2025 (UTC)

Agree with the above, would love to see some public update. Sjoerd de Bruin (talk) 19:09, 20 January 2025 (UTC)

Hey @Risker and @Sjoerddebruin! Thanks for the note of support. I've just updated the membership list.

We have had a couple of onboarding meetings where I shared information that's public about our annual plan and strategic direction. This was an important first step for members to begin discussions with the same information. This weekend is the first in-person gathering, and we'll publish what the group discussed, outcomes and next steps shortly. I'm hoping a short response before end of day, and then a full report on our materials and progress later. SDeckelmann-WMF (talk) 16:26, 26 January 2025 (UTC)

Thanks, Selena. Will look forward to hearing more. I wish the PTAC a vibrant and productive meeting. Risker (talk) 16:44, 26 January 2025 (UTC)

Agree with the above, thanks for the update. Sjoerd de Bruin (talk) 17:07, 26 January 2025 (UTC)

Meeting agenda, participation information for 24 January 2025 Product and Technology Advisory Council meeting?

Please, post the Meeting agenda, relevant supporting documentation, PTAC Member bios and participation information for Community Members for the next Product and Technology Advisory Council meeting on 24 January 2025 in Portland, Oregon.

Thank you in advance, -- Ooligan (talk) 20:44, 17 January 2025 (UTC)

Hi @Ooligan, thanks for your note. This first in person meeting was for the group to learn more about one another and to just begin to figure out how to function as a group. I left a note above to Risker and Sjoerd de Bruin giving a rough timeline for publishing our work from the last two days (and today!). We'll leave a note on this page when that's ready (soon!). SDeckelmann-WMF (talk) 16:30, 26 January 2025 (UTC)

Thank you, -- Ooligan (talk) 07:14, 7 February 2025 (UTC)

January meeting of the council

The Product and Technology Advisory Council and it's facilitators just wrapped up an intense 3 day meeting over the last weekend. After our first online meetings to get to know each other and to learn more about Product and Technology, this was a first opportunity to really dig more into what our goals and processes for the council will be and to apply ourselves. We intend to publish output of this meeting in about a week time, but we wanted to share this picture of the group today. —TheDJ (talk • contribs) 01:53, 27 January 2025 (UTC)

Woop woop! ·addshore· ^{talk to me!} 16:43, 2 August 2025 (UTC)

Draft recommendation available for feedback

We have published the recommendation from our January meeting at Product and Technology Advisory Council/February 2025 draft PTAC recommendation for feedback, and are requesting feedback at the talk page by 21 February. GorillaWarfare (talk) 00:13, 5 February 2025 (UTC)

Second draft recommendation available for feedback

Following recent community reactions surrounding two recent initiatives, the deployment of AI summaries on English Wikipedia and the recent community feedback on ToneCheck, we have published a set of proposals at Product_and_Technology_Advisory_Council/August_2025_draft_PTAC_proposals_for_feedback to help facilitate more constructive engagement between the Wikimedia Foundation and Wikimedia communities. We are requesting feedback at the talk page by 22 August. Sohom (talk) 17:34, 6 August 2025 (UTC)

Question - Linking to Unsupported Tools Working Group

Is there a reason why Product and Technology Advisory Council/Unsupported Tools Working Group seems to be an orphan page, not even being linked to from the main PTAC page? - Fuzheado (talk) 17:10, 9 October 2025 (UTC)

Re-upping this question. If there are no objections, does it make sense for someone to make a link? - Fuzheado (talk) 21:57, 20 October 2025 (UTC)

Done Sohom (talk) 15:25, 25 October 2025 (UTC)

Update to second recommendation

Based on the feedback received from the community, we have iterated on our recommendations. A updated set of recommendations which we sent to the Foundation is present at [1] Sohom (talk) 17:45, 14 October 2025 (UTC)

Future topic suggestion (meta)

Great what you're doing, thanks to all involved. I'd to suggest an important topic for some future work of the council:

• How could technical development capacities be increased?

There are many issues on phabricator and wishes in the Community Wishlist, often open for over a decade. The community and many contributors are increasingly frustrated. See for example Community Wishlist Survey 2022/Larger suggestions/1% (note: and 1% would be too little; and more funds for development are not really needed to show a campaign banner, or to show categories on mobile or to add a button that scrolls down at the top of wikidata item pages etc).

Here's a list of ideas of what could be done – e.g. as inspiration or to explain that many different things could be done or that some of these wouldn't require any extra funds. There won't be much innovation and technical progress and a healthy technical infrastructure if there's way too much technical development to even just implement solutions to urgent issues that have persisted for many 8+ years.
Please consider this topic for further deliberations, thanks.
Prototyperspective (talk) 00:42, 26 November 2025 (UTC)

@Prototyperspective (and this is my personal opinion) -- I think what you are saying has already (for the most part) there. A large portion of what folks work on is based on tasks in the community wishlist or similar. I think the main problem that needs to be tackled is engagement with the wishlist itself as a process. For what it's worth we do a fair bit of developer advocacy (I joined the movement through that process), but mentoring developers to get them started with a new codebase takes time and it is often hard for folks to tackle the really difficult problems is going to be hard problem to solve in itself. Sohom (talk) 19:52, 31 January 2026 (UTC)

I think what you are saying has already (for the most part) there. Not sure what you meant to say. But afaik none of the proposals there have been done so far. I didn't say there was absolutely no activity and efforts to get more developers and increase development capacity, just that it's by far not enough and by far not as much as it could readily be. There's lots of relatively nondifficult important issues too and the proposals are exclusive to also hiring more developers (that's actually one of the things proposed). Prototyperspective (talk) 19:57, 31 January 2026 (UTC)

@Prototyperspective I was replying to There are many issues on phabricator and wishes in the Community Wishlist, often open for over a decade. and the general spirit of this wish when I said I think what you are saying has already (for the most part) there. To my understanding there has been a significant increase in the amount of work being picked up through the Community Wishlist from WMF staff/through the annual hackathon. For what it's worth, none of the things that you propose as "not requiring extra funds" are a zero-sum game -- even something as simple as Community_Wishlist/W225 requires a fair bit of design and engineering work since it requires thinking about how your average reader (who is not a Wikipedian) might react to a sea of links showing up at the bottom of the screen. Similarly, phab:T142082 is not as simple as "add a button to add a statement" but rather "figure out what is wrong with the Wikidata interface from a usability POV and redesign it" which is a significant undertaking.

just that it's by far not enough and by far not as much as it could readily be: I get the sentiment you are expressing but there are 15 (read: a lot) variables that you are not considering here. Scaling up capacity is not as simple as "just put out a banner" (which might have high initial conversion, but I doubt any people will make it to the point to be able to solve complex bugs -- since as I explained getting developers acquainted with a codebase is expensive and requires mentoring) or "just hire more people from the global south" (WMF already hires from a variety of countries -- and in doing so needs to pay everyone roughly competitive wages in accordance with purchasing power parity, not to mention that in this context, you would also need to have engineering managers and a PM all of which are a significant undertaking) Sohom (talk) 12:54, 4 February 2026 (UTC)

From the data, it's not really significant. But let's just assume it is: a significant increase does not imply it's close to being nearly sufficient or reasonable (and that's another thing than optimal). Technical development is minimal and lots of highly important bugs remain unfixed, highly-supported issues open for many years, etc. the wish you linked does not require such design and engineering work. There usually is no sea of blue at the bottom and when there is, that's not a problem and is also the case on desktop. One could implement it in a more thoughtful way that involves more design and engineering work but that's not needed for at least showing these as links. Moreover, they could test it and/or survey instead of just "thinking" in hypotheticals. The issue you linked is as simple as that, no need to do what you suggest. Scaling up capacity is not as simple as didn't say it would be "just put out a banner" that's one low-hanging fruit to get at least started and I also wrote about landing page etc that are associated with the banner (which can partly or even fully be implemented by community volunteers) doubt any people will make it to the point to be able to solve complex bugs perfectly fine because this isn't just about complex bugs getting developers acquainted with a codebase is expensive and requires mentoring it depends on the developers; for those that need mentoring and relatively more resources, I'm arguing this should be done "just hire more people from the global south" don't know who you're quoting but it's not me WMF already hires from a variety of countries btw, is there any chart of which countries how many WMF developers are living in? significant undertaking go for it then. Prototyperspective (talk) 15:15, 4 February 2026 (UTC)

Annual plan response

We have published the consolidated response we provided to the WMF regarding the prompts asked as part of it's annual planning process on this page. Feel free to provide feedback or point out anything that we have missed or could have done better on the talk page! Sohom (talk) 05:18, 4 March 2026 (UTC)

Draft recommendations for feedback

We have published a draft PTAC recommendation on a model for affiliates interested in technical projects. Feel free to provide feedback or point out anything that we have missed on the talk page by May 8th! Sohom (talk) 19:29, 1 May 2026 (UTC)

Call for discussion topics

In our June call, PTAC agreed to establish an agenda-setting subcommittee. The purpose of this smaller working group is to better determine which topics are most appropriate for PTAC discussion. The committee has now been established and consists of Benjamin Mako Hill, GorillaWarfare, Sohom Datta, Éder Porto, Fjmustak, and Rita Ho.

While we will review WMF proposals for topics to discuss, we would also like to invite the community to suggest topics for PTAC to consider. For now, we have the following ground rules about the topics we will consider:

• The topic must be one that is not already being discussed/being resolved by the community and the WMF through other avenues (for example, Abstract Wikipedia)
• The topic must be a broad picture (i.e., it must not be about a specific feature deployment) that would benefit from discussion due to complex or diverse opinions, or sensitive topics not easily discussed in an open written forum.

We also will remain mindful that not everyone on the sub-committee is knowledgeable of topics across different language communities and projects, and will consult with the relevant PTAC members when evaluating whether those should be prioritized.

PTAC is considering two strategic topics as starting points. The first is the Reader Strategy, which means exploring and making recommendations on how we reach and engage more readers. The second is the Contributors Strategy: which types of editors we should focus on, and how to best enable them to achieve the impact we need.

We welcome suggestions on this talk page for topics you would like the committee to consider. Sohom (talk) 00:53, 31 August 2026 (UTC)

Something that's been on my mind recently: how can we better help editors at enwiki become aware of wikis in their native language or languages which they speak (also sister projects)? As it's in what's effectively the 'global lingua franca', enwiki (and to a lesser extent frwiki etc.) has a special responsibility to direct editors to smaller wikis where their participation would be much more consequential and effective at furthering the goals of the movement. Ofc some will just prefer to stay at enwiki because of its reach, but this has become an increasingly pressing issue with the use of LLMs to compensate for a poor command of the language (and I'm sure this isn't limited to enwiki). Kowal2701 (talk) 04:50, 31 August 2026 (UTC)

(suggested it for LLM PasteCheck [2]) Kowal2701 (talk) 07:40, 31 August 2026 (UTC)

I was talking to @Sthottingal-WMF yesterday at WCI 2026 and there are definitely some interesting questions especially wrt to how the WMF and the community can better promote smaller wikis (especially in the context of Google Search often de-prioritizing multi-lingual content. Sohom (talk) 05:19, 5 September 2026 (UTC)

Topic: Should the Wikimedia movement deliberately position Commons (our largest database) as a curated source of training data for open image-generation models, and on what terms? (Related: Community_Wishlist/W213)

A concrete example of what this could mean: extract a smaller, high-quality, deduplicated subset from Commons (verified licensing, authorship and semantic tagging), paired with a movement policy that asks model builders to publish the exact list of images used and to release the base model, prior to fine-tuning, openly. The aim would be a transparent public alternative to proprietary models, in line with the UNESCO Recommendation on the Ethics of Artificial Intelligence. Software Heritage (SWH) did something comparable for source code with StarCoder2 (link; also: my slides); that precedent is worth examining closely, including what worked and what can be improved.

This is a broad strategic question rather than a feature request, and as far as I know the movement (volunteers, affiliates, researchers) has no agreed position on it. It also brings together several disagreements currently being had separately: whether CC BY-SA gives the movement any real leverage over model weights, or only the soft power of a voluntary norm - one alternative would be attaching conditions to bulk access, e.g. via Wikimedia Enterprise; what contributors were promised, given that SWH offered an opt-out and free licences (CC BY-SA) arguably preclude one here; whether the curation and structured-data effort is worth its cost to Commons; how any of this interacts with generated images flowing back into Commons; etc. —super nabla(🪰 msg) 13:25, 1 September 2026 (UTC)

Further reading: https://creativecommons.org/2026/09/03/guidance-on-using-cc-licenses-in-an-ai-ecosystem/ —super nabla(🪰 msg) 11:24, 4 September 2026 (UTC)

@LWyatt (WMF) I remember seeing some intiative/deliberation in enterprise that looks vaguely similar, do you have thoughts? Sohom (talk) 05:06, 5 September 2026 (UTC)

Thanks for tagging me Sohom. As you are aware (or at least have correctly intuited) the enterprise APIs/Datasets do not currently have a specific service for Commons - we started with the text-based Wikis, then more recently added Wikidata, and are currently focused on building-out that (see our most recently quarterly tech updates). Commons is a wholly different animal and so we do not yet have specific service there. Product and infrastructure research is underway but is complex: from a question of license varieties and format diversity, commercial-scale users and usecases, metadata (captions, categories, structured-data), legal compliance (upon us, and upon reusers)... but the size and content has made it difficult to be cost-effective to work with so beyond limited partner testing we have not considered a broader product launch and folks are welcome to submit requests (there is a community-request column on our Phabricator board) that will help us look at product work in the future. Notably, the Project principles emphasise that any service/feature built will neither provide 'exclusive' access to individual paying client, nor restrict those services/features to only those who pay. For followup on these matters best to talk with CPetrillo-WMF.

From the P&T department side: there wasrecently created a media platform team, that the initial focus is on reducing risks for the platform (specifically, coming up with a new architecture for thumbnails), but also plan to think more about what responsible media reuse could look like as part of the FY 26/27 annual plan (PP3: Scalable System for Responsible Reuse). For more info on that area, please talk to BMueller (WMF).

Sincerely, LWyatt (WMF) (talk) 14:10, 7 September 2026 (UTC)

Thank you, @LWyatt (WMF) and Sohom Datta: this is a very useful picture of where things stand and of the involved groups. One clarification on what I was proposing (since I don't think it's covered by the work you describe). PP3 and the Media Platform Team address how responsible reuse could be built; my question is whether the movement should take a position on the terms of that reuse, and if so, which. Two things I'd still like the committee to consider:

1. Does CC BY-SA give us any actual leverage over how (text-, graph-, or image-based) model weights are released, or only a voluntary norm we can ask reusers to accept?
2. What does the opt-out precedent by SWH/StarCoder2 suggest to us here (including possible limitations)? Free licences (CC BY-SA) arguably preclude that opt-out in our case.

To be clear, this does not imply any preferential access for anyone (this would be a condition of use, applying to anyone who takes the data); thus, I don't think it runs counter to the project principles. Happy to expand if PTAC takes this up or to follow up w/ people mentioned above. Cheers,—super nabla(🪰 msg) 15:07, 7 September 2026 (UTC)