Proposal 1: Use local SecurePoll instance

What it says on the tin, previous versions of the arbcom election have occured on vote.wikimedia.org (colloquially called VoteWiki) with the help of WMF T&S staff. This year, with the introduction ofWP:AELECT, we now have the ability to run elections locally without WMF staff help. Should we migrate to using the enwiki instance of SecurePoll?Sohom (talk)06:19, 1 September 2025 (UTC)[reply]

1. Purely from a local governance point-of-view, I don't think it makes sense to bother WMF T&S staff about our elections especially since we can get effectively the same amount of secrecy (in theory). The one additional admin that gets added is that one-or-two members of the election-comission will hold the all important key pairs with the ability to decrypt the election instead of WMF staff.Sohom (talk)06:19, 1 September 2025 (UTC)[reply]
2. From a operations point of view, it is important to practice procedures on a regular basis to remain proficient at them. By making routine use of our local instance, we continue to build our institutional competency about how the software and processes around it work. It also gives us greater flexibility of scheduling, since we don't have to work around other demands on the singleton WMF instance. I don't see any downsides.RoySmith(talk)11:41, 1 September 2025 (UTC)[reply]

   The first two sentences of this argument don't make sense. Admin elections use the local SecurePoll instance and happen more than twice as often as ArbCom elections so even if this proposal fails we will have plenty of institutional competency with SecurePoll.* Pppery *_{it has begun...}16:38, 1 September 2025 (UTC)[reply]

3. Removes some level of external dependency and seems like a no-brainer to do so. --qedk (t愛c)14:05, 1 September 2025 (UTC)[reply]
4. I honestly can't see a reason not to hold the vote locally.Izno (talk)00:22, 2 September 2025 (UTC)[reply]
5. Like Izno, I can't think of a reason not to.QuicoleJR (talk)01:19, 2 September 2025 (UTC)[reply]
6. The current process is a hack. The fact that votewiki shows a "Log in" option in the corner that you'renot supposed to click is confusing (phab:T289542,phab:T98013). Running locally avoids this and other issues. –SD0001 (talk)08:07, 2 September 2025 (UTC)[reply]

   Also, when my changes inphab:T397565 are merged, we'll be able to entirely eliminate the process of generating voter lists, in favour of the software determining eligibility on the fly. –SD0001 (talk)08:09, 2 September 2025 (UTC)[reply]

7. I support this.^Thanks,L3X1◊distænt write◊14:31, 2 September 2025 (UTC)[reply]
8. Good idea. --LCUActivelyDisinterested«@» °∆t°22:57, 2 September 2025 (UTC)[reply]
9. Why not?Ophyrius (he/him
   T •C •G)10:24, 3 September 2025 (UTC)[reply]
10. It is fine.--Snævar (talk)12:43, 3 September 2025 (UTC)[reply]
11. Sure.fanfanboy(block_talk)17:02, 3 September 2025 (UTC)[reply]
12. This seems like a reasonable devolution of the process for simplicity's sake, now that it is technically feasible.SnowRise ^{let's rap}22:08, 3 September 2025 (UTC)[reply]
13. Let us rely on WMF Staff where we genuinely need them. This will also future proof our polling ability with more resilient tooling ~ 🦝Shushugah (he/him • talk)17:35, 4 September 2025 (UTC)[reply]
14. T&S are busy people and AELECT has shown that we could do ourselves. This would also remove an item from the global VoteWiki calendar, which does not handle scheduling conflicts well (though the exact constraints are not as simple as "elections can't overlap" iirc).Toadspike[Talk]09:29, 7 September 2025 (UTC)[reply]
15. Quite obviously a step in the right direction. Not only for the flexibility in timing that it affords, but also towards relieving the burden on the WMF to run it.Kudpung กุดผึ้ง (talk)11:29, 7 September 2025 (UTC)[reply]
16. We shouldn't intrude on time (=donor money) from the WMF more than necessary.HouseBlaster (talk • he/they)21:33, 7 September 2025 (UTC)[reply]
17. Absolutely. This project is larger than many organizations and entities that conduct their own elections; there is no reason, now that the ability has been developed, for us not to manage our own election processes entirely within this project.Risker (talk)22:02, 7 September 2025 (UTC)[reply]
18. This would make the ArbCom election process easier for both us and the WMF.ModernDayTrilobite (talk •contribs)14:06, 8 September 2025 (UTC)[reply]
19. Doing this has several advantages. You wouldn't need to schedule the election with the WMF to avoid collisions with other elections (and the language settings needed for each election). Additionally, it would not require a user to visit another wiki to vote in an election (though this is largely transparent for votewiki). The admin elections proved that the system can work when it's just on the English Wikipedia.DreamyJazz^{talk to me |my contributions}20:26, 8 September 2025 (UTC)[reply]
20. Seems just clearly easier this way.Galobtter (talk)02:24, 10 September 2025 (UTC)[reply]
21. Support in principle, though per the discussion below this needs to be undertaken carefully. Multiple election admins should be required, and it should be required that they acknowledge redundancy in key management and availability throughout the entire election cycle. —xaosflux^Talk13:25, 10 September 2025 (UTC)[reply]

    It probably would make sense to have WMF staff keep copies of the keys as a disaster recovery hedge.RoySmith(talk)13:33, 10 September 2025 (UTC)[reply]

22. Increases flexibility in election timing, reduces dependency on WMF staff, and slightly improves user experience for voters (not having to visit a different site).the wub"?!"23:28, 10 September 2025 (UTC)[reply]
23. SupportFaviFake (talk)18:52, 13 September 2025 (UTC)[reply]

1. As a separation of powers matter I would prefer that the enwiki power structure does not control a vote to determine the enwiki power structure.* Pppery *_{it has begun...}16:36, 1 September 2025 (UTC)[reply]

   Does administering the vote via VoteWiki actually separate power? What is really the concern here? The people with access to the votes lying about the results?Czarking0 (talk)23:18, 1 September 2025 (UTC)[reply]

   External stewards still have to certify the tally. –SD0001 (talk)07:59, 2 September 2025 (UTC)[reply]

   Under this proposal, yes. A potential additional proposal that has already been murmured about below would remove the stewards. (I am favor of doing so.)Izno (talk)16:14, 2 September 2025 (UTC)[reply]

2. I'm with Pppery on this; I'd definitely prefer ACE to have at least some level of oversight from the Foundation, and isolating the vote to votewiki is a good way of doing this.stwalkerster (talk)17:48, 6 September 2025 (UTC)[reply]

   I don't think WMF scrutiny will be going away,WP:ARBCOM election is typically pretty closely followed by Wikimedia Foundation staff folks. What will be going away is our explicit dependence on them.Sohom (talk)04:29, 7 September 2025 (UTC)[reply]

3. What Pppery said, to some degree I agree with. I also dislike the fact that the election clerk user-group can only be assigned to current admins, meaning that a non-admin would face even more challenges in submitting themselves to be on the electoral commission than already exists. The current setup with the WMF configuring SecurePoll and stewards keeping things running on votewiki seems fine already.EggRoll97^(talk)23:28, 6 September 2025 (UTC)[reply]

   Note this proposal isn't proposing a change to the role of the electoral commission, and thus doesn't require them to be election clerks for a locally hosted poll.isaacl (talk)04:22, 7 September 2025 (UTC)[reply]

   Based on the original proposal and what Sohom commented below, itis in fact changing the role of the election commission, or it was originally prior to more discussion. My main objection of course is going to be that the election clerks are local admins exclusively and that there are no non-admins as election clerks, meaning there is a more limited pool of people administering the technical side of the elections.EggRoll97^(talk)07:58, 7 September 2025 (UTC)[reply]

   Limiting the election clerk role to administrators was just something that the limited set of participants atWikipedia talk:Administrator elections/SecurePoll permissions proposal § Make assignable by any sysop? compromised on to move the process forward. If there is sufficient interest, the community can establish a consensus for a different process to grant the election clerk role. (I believe the outcome of the discussion between Sohom and me is that the proposal does not require electoral commissioners to be election clerks.)isaacl (talk)17:14, 7 September 2025 (UTC)[reply]

   That, election commissioners don't need to be election clerks, the election clerks are community volunteers with the election clerk right. As isaacl said, if there is significant interest the community can have a consensus process for electing/giving out election clerks. The admin-only restriction was a quick and dirty solution for AELECT.Sohom (talk)18:17, 7 September 2025 (UTC)[reply]

   The election admins do nothave to be part of the electoral commission. ELECTCOM's primary role is decision making, not paperwork. Other election volunteers could be election admins (perhaps with the approval of electcom). —xaosflux^Talk10:29, 9 September 2025 (UTC)[reply]

• Why is this desirable?Graham11 (talk)06:22, 1 September 2025 (UTC)[reply]
• Does this proposal include keeping stewards as scrutineers, or does it propose having community members do scrutineering? Is there a process proposed for selecting election clerks?isaacl (talk)06:33, 1 September 2025 (UTC)[reply]

  Election clerks would be the same as the folks on the election-commission (as it has always been). The proposal assumes stewards will still be scrutineers, tho that could change if this proposal gains enough momentum and there is interest in that being changed.Sohom (talk)06:41, 1 September 2025 (UTC)[reply]

  No, the original purpose of the election commission's role was to expeditiously resolve issues which couldn't wait for a community consensus discussion, not to be in charge of running the election. The election continues to be run by community volunteers (which can overlap with the commissioners, but is not required). I think there is value in keeping the commissioners uninvolved with the mechanics of the election, in order to preserve more of an outside viewpoint should they need to address procedural issues.isaacl (talk)06:48, 1 September 2025 (UTC)[reply]

  Fair, enough, maybe we have a expanded role of the election commission with two technical volunteers "running" the election and 3 volunteers in the more traditional role of commissioners who are able to resolve issues that cannot wait for community consensus?Sohom (talk)07:07, 1 September 2025 (UTC)[reply]

  I think it's easier to adopt the same terminology as for administrator elections: have two election clerks in charge of managing the actual poll, replacing the WMF staffer. The rest of the election process can continue to be managed by community volunteers, and the election commission can continue with its current scope.isaacl (talk)07:30, 1 September 2025 (UTC)[reply]

  Yes, that makes sense!Sohom (talk)07:48, 1 September 2025 (UTC)[reply]

  Perhaps using local checkusers instead of stewards should be a separate question, to make things easier for the closers. –Novem Linguae(talk)07:07, 1 September 2025 (UTC)[reply]

  That was my thought-process as well.Sohom (talk)07:07, 1 September 2025 (UTC)[reply]

  I raised these questions now because there isn't a lot of time for another RfC afterwards to resolve them. The experiences we've had with administrator elections makes me wary of assuming that the community can quickly come to agreement.isaacl (talk)07:26, 1 September 2025 (UTC)[reply]

• Has the staff asked that we do this? This seems like we could applyIf it ain't broke don't fix it.Czarking0 (talk)06:36, 1 September 2025 (UTC)[reply]

  Staff has not asked us to do this (yet). However, at a logistical and technical level, while it might not be totally broken, but it definitely less than ideal. The number of people who have hands-on experience with running elections throughSpecial:SecurePoll currently is less than 10 people (actually the number is closer to 4-5 people if you consider T&S staff and volunteers who ran theWP:AELECT elections). More importantly, VoteWiki can at a time only support a single wiki holding a election since multiple moving parts need to be configured properly (one particular configuration for example being the localization of the wiki). Currently, communities uses VoteWiki to conduct their votes (alongside global elections like the U4C), this can (and often does) create scheduling issues. I think it makes sense for us to build the capability within the community to run our own elections so that we can stop depending on a process that is both messy and not very well supported.Sohom (talk)07:02, 1 September 2025 (UTC)[reply]

  I agree with the value of spreading the knowledge of how to run secure elections using SecurePoll on the enwiki server. My understanding, though, is that the limitation on multiple elections on the VoteWiki server is due to WMF personnel availability, and not an inherent technical restriction. The VoteWiki server itself can only be localized to one particular language at a time, but the SecurePoll software is designed to hold a poll that is translated into multiple languages, so I believe it should be able to handle different languages for different polls (I appreciate that other localization aspects like number formats probably haven't been tested much, if at all).isaacl (talk)07:17, 1 September 2025 (UTC)[reply]

  SecurePoll (the extension) has been translated to multiple languages and be used in multiple language simultaneously, but I've seen mentions of having to change the language of votewiki during elections (for example, seephab:T361902#9820281) when preparing the wikis for non-english communities. This might be purely aesthetic, but it is clearly part of the protocol WMF staff follow (at times?) and these do add overhead to the process that would not be present if this was run on enwiki.Sohom (talk)07:40, 1 September 2025 (UTC)[reply]

• WP:AELECT2 kind of blazed the trail for holding local elections. There were no problems. I documented everything atWikipedia:Administrator elections/July 2025/SecurePoll setup. It made a lot of sense for AELECT since T&S, stewards, etc. are willing to do about 1 election a year, and AELECT wanted to have 2 elections a year. Both groups keep pretty busy and elections aren't their main thing. However since ACE is only once a year, it is not as pressing for ACE to switch. Up to you all. –Novem Linguae(talk)07:12, 1 September 2025 (UTC)[reply]

  You're probably more familiar with SecurePoll than most other editors, so I'll ask you: would there be any problems holding AELECT and ACE at the same time, in terms of SecurePoll configuration? Or can it easily handle simultaneous elections? --rchard2scout (talk)12:21, 1 September 2025 (UTC)[reply]

  Nope. No (technical) problems with simultaneous elections locally. It's very scalable. AELECT3 is currently scheduled for december, but we are going to RFC that to double-check. Because there may be some non-technical reasons (voter fatigue) not to overlap AELECT and ACE. –Novem Linguae(talk)17:48, 1 September 2025 (UTC)[reply]

• I'd think the biggest risk here is that the volunteer(s) mess up the encryption key and render the entire election ruined. Strong key handling procedures will need to be followed. —xaosflux^Talk18:48, 5 September 2025 (UTC)[reply]

  Straying off topic, butUser:SD0001, you mentioned a patch above, I don't think the patch would meet our suffrage requirements - meaning we will still need a manual list (or someone will quickly need to suggest changing our requirements). Am I missing something? —xaosflux^Talk13:00, 7 September 2025 (UTC)[reply]

  No, I'm seeing the same thing. We'd need to generate the list outside of SecurePoll, under the current suffrage rules. The automated checks would be insufficient. That being said, generating them with my script is not an issue and it generates the list quite fast. I'm happy to do it again this year, if needed. —CYBERPOWER(Message)22:07, 7 September 2025 (UTC)[reply]

  I think the point here is not "it will immediately be operational" but rather "it will make it a lot easier for us to request the specific knobs if we want".Sohom (talk)23:24, 7 September 2025 (UTC)[reply]

  If the patches are merged in time for the election, a manual list can be avoided. I'm not a fan of the continued use of non-open code for such a crucial part of the process. –SD0001 (talk)02:54, 10 September 2025 (UTC)[reply]

  The code is published (seeWikipedia:Arbitration Committee Election/Coordination § Tools), and the resulting lists are provided publicly before being entered into the poll configuration so anyone can double-check them. (I do think everyone agrees that automated checking of the eligibility criteria is ideal.)isaacl (talk)05:08, 10 September 2025 (UTC)[reply]

• As a previous electoral commissioner that has served multiple times, I am closely watching this proposal. Xaosflux is very much correct that mishandling keys can definitely cause some major disruption. I think there should be some discussion on proper key management, else the EC will just proceed at their own discretion.—CYBERPOWER(Message)21:56, 7 September 2025 (UTC)[reply]

  @Cyberpower678 If I understand correctly there are two keys, one public, one private that can be generated using openssl,Wikipedia:Administrator_elections/July_2025/SecurePoll_setup documents the exact commands to run, once they are generated, they need to be among the two election clerks and must not be deleted until after the certification of the election (during which they will needed to be inputted into the system). If I remember correctly, SecurePoll will scream at you if do put in the keys in the correct format, so the chances of "messing up" imo are less than we are making it out to be. The only thing to keep in mind is tonot delete them. The election clerks can obviously run "dry-run" elections to become familiar with the system and it's intricacies (which is something that was not possible on VoteWiki).Sohom (talk)23:23, 7 September 2025 (UTC)[reply]

  One can simply accidentally replace just a single character and that's all would be needed to kill the election. Unless of course, the only course of action one can take after adding a key is deleting them. But if they can be edited as easily as I edit this page, that still can be a problem for less technical users. —CYBERPOWER(Message)23:37, 7 September 2025 (UTC)[reply]

  One thing I'd like to see ismore people involved in the process. Handling crypto keys isn't rocket science (he says withHidden Figures streaming in another window) but it's also something that takes a bit of experience to not mess up (he says with plenty of messing up experience of his own). It would be awkward if we had one person who knew how to do it and they were victimized by some errant mass transit.RoySmith(talk)00:02, 8 September 2025 (UTC)[reply]

  One can simply accidentally replace just a single character and that's all would be needed to kill the election. The OpenSSL command line commands generate files. I never edited any keys duringWP:AELECT2. It was a lot of copy pasting from files into SecurePoll text boxes, and drag n dropping the files to @Robertsky on Discord (who held a copy of the keys as an insurance policy against losing the keys / bus factor). I suppose your hypothetical about accidentally selecting one character, accidentally replacing it with another character via keyboard typo, and then accidentally saving the file, all before the keys were transferred to the backup person is possible, but it would require those 4 things to happen in a row. –Novem Linguae(talk)03:23, 8 September 2025 (UTC)[reply]

  (Note that I misspoke here, there are two key-pairs, so four keys in total), one public for encryption, one private for signing need to be put in when setting a election up, the keys are editable until the elections start but are then frozen once the election starts. To tally/end the election the private key for encryption and the public one for signing must be input. To be clear, (and to your point) I am not saying it's impossible for folks to mess this up, but it is (imo) harder to mess up if you do follow the rules. The keying process should become clearer and easier the more institutional knowledge we have of this system and we do obviously need to start somewhere.Sohom (talk)00:24, 8 September 2025 (UTC)[reply]

• Given the current state of the discussion for this proposal, interested persons may want to start thinking about who they think has the desired skills and level of trust to be election clerks. As a point of reference, in the most recent admin election, Novem Linguae, Robertsky, Dbeef, RoySmith, Zzuuzz, and Dreamy Jazz were assigned as election clerks; the latter four were added to be scrutineers. (If it matters to anyone, Novem Linguae now has access to the database where all English Wikipedia data is stored, including the votes made in SecurePoll. As alluded to, the votes are encrypted, but at least two election clerks will hold the decryption key. It is against WMF policy of course for anyone with database access to access the voting data directly from the database.)isaacl (talk)15:31, 10 September 2025 (UTC)[reply]

Proposal 2: Use Local scrutineers

There has been discussion in proposal 1 regarding whether or not Stewards are still going to be scrutineers for this election, so I'm putting it as a formal proposal: If proposal 1 passes, should we use local CUs rather than stewards to scrutineer this election?All the Best --Chuck^Talk16:43, 2 September 2025 (UTC)[reply]

1. Yes, I'm pretty sure I support this. I find the concern about local CUs being abusive during scrutineering not quite in the realm of absurd. If we really must, I'd support our current count of scrutineers plus a steward just to double check over their work and/or consider the relevant CUs' potential votes themselves, but I don't see this as particularly necessary.Izno (talk)16:57, 2 September 2025 (UTC)[reply]
2. I find arguments regarding optics (or how something 'appears' to be corruption/bad/etc) quite unconvincing personally. We should be preventing things from happening because they actually have the ability to cause harm. But of course this is an unpopular opinion.dbeef [talk]20:47, 5 September 2025 (UTC)[reply]

   I also find the optics pieces that happened around some Admin election stuff unpersuasive. I obviously do not think this is merely optics as someone who has been a CU, Arb, and ACE Electcom member. Best,Barkeep49 (talk)21:14, 5 September 2025 (UTC)[reply]

3. There is no reason whatsoever for the largest project, with quite literally tens of thousands of regular users, and the largest cadre of checkusers within the movement, to be relying on people outside of this project to make sure there was no socking during voting. (Seriously, that's all that scrutineers really do. Been there, got the t-shirt.) It is incredibly poor use of steward time. Do we really have so little confidence in the ability of checkusers to act honestly? Do we really have such a poor understanding of the SecurePoll voting process that we honestly think that a couple of local CUs could directly determine the outcome of the election in a way that they favour? If that is the case, then we should return to public, visible voting for candidates and stop using SecurePoll. We have seen the enemy and they is us.Risker (talk)21:35, 7 September 2025 (UTC)[reply]
4. I was just talking to a few folks and realized that the threat model of what we are doing here doesn't make sense. There are exactly two scenarios we are defending against, firstly, our three scrutineers CUs collude (somehow, in a offwiki location?) to strike an exactly arbitrary number of votes (note that they will not know who voted for whom) to elect their prefered candidate to ArbCom. This is just a mathematically and practically impossible task even without considering the fact that striking votes is visible to the public. Secondly, it is possible that a CU logs all of the data provided by SecurePoll and then subsequently uses that data to fish for sock accounts. This (imo) is a more plausible scenario of abuse, but it is important to keep in mind that Stewards scrutineers already have the same problem and any such fishing attempt would be somewhat detectable in the logs by other CUs and the currently sitting ArbCom (since SecurePoll's scrutineering is typically not tied to the CU extension and uses a seperate logging system, checking the data against CU data would require a CU action). --Sohom (talk)23:42, 7 September 2025 (UTC)[reply]
5. I land here for myriad reasons. First, I am really struggling to see how the current setup makes things better. The scrutineers should obviously not becandidates for ArbCom. I also understand that the scrutineers should not be incumbents, regardless of if they are running for reelection or not. But the stewards already are appointed by ArbCom for the task of scruntineering, andWP:GRP is clear that without this grant they are not permitted to use CU in the normal course of their duties. We would actually bedecreasing the amount of ArbCom involvement in ACE, given that they would no longer need to appoint stewards as local CUs.

   When it comes to the threat model which Sohom alludes to, I can see one possibility they do not mention: that ArbCom would pressure the CUs to allow some socking to fly under the radar. To which I say, well, call me an optimist, but I trust the members of the CU team to report any obvious shenanigans to the community promptly and I trust arbs—people who have spent a great deal of time helping Wikipedia—to not do this.

   Steward time is very valuable. They are responsible for dealing with cross-wiki abuse and helping communities whoare too small to be self sufficient. I don't think we should consume their time more than necessary.

   Finally, on the positive side, let's look at the results of local CU scrutineering for AELECT. They got it done intwo days. Two days! More on-boarding time for the elected candidates is a great thing, less candidate stress from the wait is a great thing, more local accountability for scrutineers is a great thing.HouseBlaster (talk • he/they)00:44, 8 September 2025 (UTC)[reply]

6. Per Risker, whose argument about it being pretty much impossible for CheckUsers to influence an election without detection is persuasive. You'd have to get all 3 scrutineers in on it, guess accurately that certain people voted for certain candidates, then strike a large number of votes. This appears to be more an optics thing than an actual threat. Risker's argument about steward time and enwiki using its significant resources to support itself is also persuasive, since this frees up more time for smaller wikis to receive attention, which is equitable. Finally, I find the English Wikipedia checkusers trustworthy, and I find arguments that they may become untrustworthy in an election of their supervisors to be unpersuasive. –Novem Linguae(talk)03:30, 8 September 2025 (UTC)[reply]
7. Per above, and honestly, if we really don't trust our CheckUsers, then there is a much bigger problem at play than election scrutineering.fanfanboy(block_talk)12:39, 8 September 2025 (UTC)[reply]

1. * Pppery *_{it has begun...}16:48, 2 September 2025 (UTC)[reply]

   Reaffirming my oppose since consensus is starting to swing the other way; I'm not convinced (and probably won't be convinced) - I still feel that, quoting my comment above,As a separation of powers matter I would prefer that the enwiki power structure does not control a vote to determine the enwiki power structure* Pppery *_{it has begun...}03:52, 8 September 2025 (UTC)[reply]

2. Arbitrators are checkusers' supervisors. The Stewards are a completely independent group. I like that seperation of powers. This for me is qualitatively different than what happens in admin elections. Best,Barkeep49 (talk)18:04, 2 September 2025 (UTC)[reply]
3. Per Barkeep49.~delta^{(talk •cont)}21:43, 2 September 2025 (UTC)[reply]
4. I have the same opinion as Barkeep49.~Gwennie🐈｟💬 📋｠22:58, 2 September 2025 (UTC)[reply]
5. I think Barkeep49 is right on this one. --LCUActivelyDisinterested«@» °∆t°22:58, 2 September 2025 (UTC)[reply]
6. Also opposing per Barkeep. Using Stewards ensures better separation of powers.QuicoleJR (talk)02:04, 3 September 2025 (UTC)[reply]
7. per Barkeep49.Ophyrius (he/him
   T •C •G)10:23, 3 September 2025 (UTC)[reply]
8. My initial feeling was "Of course we should use local CUs as scrutineers", but Barkeep (as usual) makes a good point and upon reflection I agree with him. Full disclosure: I'm an enwiki CU and was a scrutineer for the most recentWP:AELECT.RoySmith(talk)13:24, 3 September 2025 (UTC)[reply]
9. Mind you, on the one hand, I think Steward involvement is probably superfluous. While I have serious concerns about how much power has accrued into ArbCom as a body, I don't view the kind of intentional and malignant interference with the election process that would have to take place to necessitate Steward involvement as at all likely (or feasible to get away with). Our community has its issues, but that sort of blatant corruption has, we can happily say, never really been present in our elections. And I'm confident in the ethics of our CUs and the relative openness of our processes to make any effort in that respect something that the community would almost immediately become aware of. But all of that said, what's the harm in having an extra layer of oversight here, really? We are taking about the selection of our most highly placed community members and arbiters of our administrative apparatus. Why not have an extra layer of independent oversight assuring the integrity of the process?SnowRise ^{let's rap}22:15, 3 September 2025 (UTC)[reply]

   However, note that ideally I'd like to see both local CUs and Stewards involved in the process, and I'm somewhat surprised this is being presented as an either/or scenario.SnowRise ^{let's rap}22:18, 3 September 2025 (UTC)[reply]

10. Circular supervision is generally a bad idea.Stifle (talk)08:12, 4 September 2025 (UTC)[reply]
11. I am not genuinely concerned, but separation of power is a model that enwp should lead by example, for other wikis as well ~ 🦝Shushugah (he/him • talk)17:35, 4 September 2025 (UTC)[reply]
12. Per Barkeep49 (separation of powers). --Minoa(talk)20:57, 4 September 2025 (UTC)[reply]
13. If a trusted external party is available, I'm inclined towards this separation of powers. Barkeep & RoySmith, with their expertise, are convincing voices. ~Pbritti (talk)21:39, 4 September 2025 (UTC)[reply]
14. Absolutely not. I don't mind local CUs providing input where requested, but decisions should be kept to a group which is independent of enwiki. The steward team mostly meets that goal, with the exception of a handful of stewards who are active on enwiki. This isn't part of the proposal, but non-local CUs (from larger projects such that they are more likely to have sufficient experience) would likely also meet that goal.stwalkerster (talk)17:59, 6 September 2025 (UTC)[reply]

    That is the reason we avoid using our steward capabilities on our homewiki's, e.g. I'd never be the scrutineer from the stewards team here. —xaosflux^Talk20:45, 6 September 2025 (UTC)[reply]

15. As a CU and former ArbCom member, I would much prefer the CU team not be involved in ArbCom elections as we are overseen by the very same people. It's important for both the integrity of the process and for the CU team to be neutral in such matters.Mkdw^talk21:25, 6 September 2025 (UTC)[reply]
16. Considering ArbCom has influence over CUs (via being able to revoke their checkuser group), I would prefer that an independent body, in this case Stewards, continue to scrutineer if possible.EggRoll97^(talk)23:30, 6 September 2025 (UTC)[reply]
17. Per most of the comments above.Thryduulf (talk)08:06, 7 September 2025 (UTC)[reply]

    Re-affirming my opinion in light of the newer support !votes. The appearance of propriety matters just as much as actual propriety. Checkusers scrutineering the election for the group who are simultaneously their bosses and close colleagues lacks all appearance of propriety because it lacks the actuality of propriety.Thryduulf (talk)04:09, 8 September 2025 (UTC)[reply]

18. Though if the Stewards complain that they are overstretched, I would consider allowing a mixed CU+Steward team.Toadspike[Talk]09:32, 7 September 2025 (UTC)[reply]
19. Independent oversight is essential.Kudpung กุดผึ้ง (talk)11:35, 7 September 2025 (UTC)[reply]
20. Barkeep says it perfectly. ~Amory(u •t •c)16:47, 7 September 2025 (UTC)[reply]
21. Per Barkeep49 re: separation of powers.Some1 (talk)19:47, 7 September 2025 (UTC)[reply]

    I read through the new support !votes and don't find them convincing or persuasive. "We should automatically trust our local checkusers" is a poor argument so I'm not going to touch on that. The scenarios presented, albiet unlikely to occur, are still scenarios that could potentially happen. And to me, ensuring the integrity of the election is more important than getting the results out a couple of days sooner.Some1 (talk)12:12, 8 September 2025 (UTC)[reply]

22. Per Barkeep,although another option would be a completely independent local body (maybe election clerks, assuming they are separate from scrutineers) that is temporarily granted CU rights for the duration of the election.ChaoticEnby (talk ·contribs)21:57, 7 September 2025 (UTC)[reply]

    One of the problems with that is that CU has a steep learning curve. New CUs are strongly urged to not use their new powers until attending a training session which covers both policy (i.e. when you are allowed to run a check) and practice (i.e. how to make sense of all the raw data the CU tool throws in your direction). You don't just hand them a secret decoder ring and send them out to play with their new toy.RoySmith(talk)22:10, 7 September 2025 (UTC)[reply]

    Indeed. There isn't an option where it's "could be anyone" doing the lookthrough.Izno (talk)22:30, 7 September 2025 (UTC)[reply]

    Agreed, we should not be sending out new/inexperienced folks for this. --Sohom (talk)23:42, 7 September 2025 (UTC)[reply]

    From what I understand, stewards are temporarily granted temporary CU rights on en.wiki for this purpose – would it be possible to do the same thing with local users who have had familiarity with CU tools before (like, CU on other projects, or former en.wiki CU) for that matter?ChaoticEnby (talk ·contribs)00:29, 8 September 2025 (UTC)[reply]

    Assuming they're currently signatories to the relevant NDA, theoretically.Izno (talk)00:39, 8 September 2025 (UTC)[reply]

    The idea of temporary checkusers that aren't stewards seems a little iffy (due to creating a fast track process to become checkuser). Blending checkuser and election clerk duties also seems a little iffy (from a separation of powers perspective). –Novem Linguae(talk)03:44, 8 September 2025 (UTC)[reply]

    Struck my vote due to significant opposition.ChaoticEnby (talk ·contribs)03:51, 8 September 2025 (UTC)[reply]

    I think it'd still be reasonable to oppose per Barkeep. I think people who replied here were just saying be careful of theanother option part. Up to you of course, but please don't feel pressured to completely remove your !vote if you are still persuaded by Barkeep's reasoning. –Novem Linguae(talk)03:58, 8 September 2025 (UTC)[reply]

    I struck the whole vote as I wasn't sure whether the added cost on stewards is worth avoiding circular supervision, and, given that my proposed alternative had issues I hadn't thought about, I preferred to take a bit more time to weigh the possibilities. After giving it some more thought, while I do believe that current CUs and arbitrators are trustworthy enough to not pull these kinds of election shenanigans, I prefer on principle to have a robust election system that minimizes single points of failure, and that does in my opinion justify a relatively low added cost for stewards.ChaoticEnby (talk ·contribs)13:03, 8 September 2025 (UTC)[reply]

23. No, not so long as checkusers serve at the pleasure of the arbitration committee. Would revisit if that role ever divests from arbcom. —xaosflux^Talk13:07, 8 September 2025 (UTC)[reply]
24. My thinking is largely analogous to Shushugah's here—to wit, I don't think there's any realistic likelihood of malfeasance occurring, but even so I think it's worth preserving the separation of powers as a matter of principle. Avoiding even theappearance of impropriety is valuable for protecting trust in the electoral system. This being said, I also sympathize with HouseBlaster's point (in the support section) about the stewards' time being limited; if the stewards themselves indicate that they would prefer enwiki CUs to take over the scrutineering, then I wouldn't object to that.ModernDayTrilobite (talk •contribs)14:01, 8 September 2025 (UTC)[reply]
25. Because CheckUsers are supervised by ArbCom, even if the process is actually independent, there is a question of optics. Your average person cannot verify that the process is independent or be able to disprove allegations of the process not being independent. For example, someone could claim that a CU was told "do this or I'll remove your CU rights" without evidence but it would be hard for an average user to verify that this claim is incorrect. With stewards doing the supervision of the election, it would be easy to disprove this.DreamyJazz^{talk to me |my contributions}20:18, 8 September 2025 (UTC)[reply]
26. It doesn't seem unaffordable to have stewards help, especially since we're the biggest wiki, and it doesn't hurt. I would be ok with both local CUs and stewards supervising, of course.Mrfoogles (talk)00:06, 9 September 2025 (UTC)[reply]
27. Per Barkeep. This just makes sense to me.Hey man im josh (talk)16:48, 9 September 2025 (UTC)[reply]
28. Separation of power and independent oversight is crucial in my opinion as (as Barkeep already said) ARBCOM is responsible for controlling them (so to speak they are somewhat dependent on an ARBCOM that is well-disposed against them)Squawk7700 (talk)21:18, 20 September 2025 (UTC)[reply]

• I'm Neutral on this proposal, As having CU's scrutineer the election that appoints CUs isn't the best, but making this election fully local is an appealing option.All the Best --Chuck^Talk16:43, 2 September 2025 (UTC)[reply]
• Might want to add an "If Proposal 1 passes" to the wording of this RFC, to make it clear that it's conditional on Proposal 1 passing. I don't think having enwiki checkusers scrutinize a votewiki election is what is intended with this proposal (although correct me if I'm wrong). –Novem Linguae(talk)19:42, 2 September 2025 (UTC)[reply]

  @Novem Linguae, that is correct. I'll update accordingly.All the Best --Chuck^Talk21:20, 2 September 2025 (UTC)[reply]

  Although it might not have been what you intended, personally I don't think where the poll is hosted should matter. In either case, the considerations of trust and separation of responsibilities are the same.isaacl (talk)01:01, 3 September 2025 (UTC)[reply]

  The reason "If proposal 1 passes" is important, in my opinion, is if 1 were to fail and 2 were to be consensus for using enwiki checkusers. That would create a situation where we'd try to get enwiki checkusers to scrutinize a votewiki poll, which is a weird and complicated outcome. –Novem Linguae(talk)04:55, 3 September 2025 (UTC)[reply]

  It would be a new approach, but personally I don't think it would be particularly weird or complicated. The step of the arbitration committee granting checkuser access to the scrutineers would become unnecessary.isaacl (talk)09:13, 3 September 2025 (UTC)[reply]

• Looks like the RFCs are currently leaning towards local elections + steward scrutineers. Just a note that this combination will require special approval from ArbCom every year since they'll have to appoint the stewards as temporary enwiki checkusers through a special process. Someone should notify ArbCom in advance so they are not taken by surprise. –Novem Linguae(talk)22:17, 3 September 2025 (UTC)[reply]

  The arbitration committee already grants temporary checkuser privileges each year to the steward scrutineers for the arbitration committee election, so there should be no surprises.isaacl (talk)22:36, 3 September 2025 (UTC)[reply]

  Asking ArbCom to do so is also on theACE checklist. Best,Barkeep49 (talk)23:00, 3 September 2025 (UTC)[reply]

  Yes; this year I made areference task list and used to startthis year's list of tasks. (Ialso alluded to this step in an earlier comment.)isaacl (talk)00:24, 4 September 2025 (UTC)[reply]

• Netural, I have been on the fence about this (and continue being on the fence about it). I think the argument for "appearance of a seperation of powers" is a strong one, but if I think "how likely is this to happen in the real world", the only number I can convincingly come up with is "very very (very) unlikely" (yes, I know that is not a number). --Sohom (talk)16:02, 5 September 2025 (UTC)[reply]
• I don't really understand the process in enough detail to be confident voting, as I'm not really sure how significant the risk of running it on enwiki is if the election committee is ?corrupt? or something. However, I would like to mention that running both elections on the same system would probably be significantly better, and I suspect not having to change websites may reduce friction and possibly increase voting.Mrfoogles (talk)00:11, 9 September 2025 (UTC)[reply]
• It is inconceivable to me that arbcom (as a body or as individuals) would try to influence the outcome of an election. It is equally inconceivable to me that local CUs acting as scrutineers would allow such interference. Yet one must only read the newspapers to discover that things which were previously thought to be inconceivable are happening every day. It is encumbent upon us to imagine inconceivable things happening and build systems which are as resistant as possible to those eventualities.RoySmith(talk)11:23, 9 September 2025 (UTC)[reply]

Proposal 3: Modify confidentiality requirements to be part of nomination statement

Candidates are required to verify that they have completed and submitted theConfidentiality agreement for nonpublic information upon posting their candidacy statements. This is similar to the requirement that they confirm alternate accounts in their candidacy statements, and similarly should not be part of the "word count" for the statement.Risker (talk)14:22, 7 September 2025 (UTC)[reply]

1. Completion of this agreement is mandatory for elected candidates to be inducted into the Arbitration Committee. Any delay in completion adversely affects the entire group of elected candidates when initiating onboarding and orientation. No non-public information is required to complete the agreement. Note that election monitors/scrutineers/managers are not required to verify that this document has been submitted; that responsibility lies with the person or persons who grant access to the newly elected arbitrators.Risker (talk)14:22, 7 September 2025 (UTC)[reply]
2. Support. I think this is a good idea. I'm not aware of any downside to signing the agreement without having any reason to do so, meaning that unsuccessful candidates are not disadvantaged in any way.Thryduulf (talk)14:43, 7 September 2025 (UTC)[reply]
3. * Pppery *_{it has begun...}14:52, 7 September 2025 (UTC)[reply]
4. Seems like a good idea to me.Some1 (talk)19:45, 7 September 2025 (UTC)[reply]
5. I conditionally support this, as long as candidates are aware of this requirement during the nomination process, it is clarified during the process that no non-public information is required to sign the agreement, and the process of doing this is streamlined. --Minoa(talk)21:28, 7 September 2025 (UTC)[reply]

   Vote has been modified to conditional support pending the streamlining of the process. --Minoa(talk)22:24, 10 September 2025 (UTC)[reply]

   Can you clarify what aspects of thesigning process you feel should be streamlined? Note, though, it's a WMF process so there isn't much the English Wikipedia community can do about it.isaacl (talk)23:17, 10 September 2025 (UTC)[reply]

   Good question. My modification of the vote was in response to concerns by xaosflux and I thought, how can we streamline the process further? --Minoa(talk)00:40, 11 September 2025 (UTC)[reply]

6. Makes the transition less hectic.—Femke 🐦 (talk)20:26, 8 September 2025 (UTC)[reply]
7. Given that if they are elected they would need to do this already, I think it's best if this is done in advance to avoid any surprises. Ditto with Risker's thoughts.DreamyJazz^{talk to me |my contributions}20:27, 8 September 2025 (UTC)[reply]
8. stwalkerster (talk)13:17, 10 September 2025 (UTC)[reply]
9. Support.Jcoolbro (talk)(c)15:19, 10 September 2025 (UTC)[reply]
10. Seems reasonable. There's still time for candidates to do this.Travellers & Tinkers (talk)15:33, 12 September 2025 (UTC)[reply]

1. This is fixing a problem that doesn't exist. We already require users to promise to sign, and in a very long time it hasn't been an issue (or if it has, it was "will you sign" at the discussion and when that was a "no" the candidate was rightfully rejected if memory serves). And contrary to the suggestion below that the arbs who don't want the additional paperwork before running are probably not going to be good arbs, I was perhaps one of the top 3 active users on the term I sat. Fundamentally, addinganother requirement to what is the already onerous process of getting people to run is not something we should be supporting, when we already have pain and misery even getting enough candidates.Izno (talk)16:23, 7 September 2025 (UTC)[reply]
2. I can see why this was proposed, but to me it seems unnecessary, I think the current "hey you should promise to do this" with a heavy nod towards doing it before getting elected is fine enough. If there are folks who are putting it off until the last momen, that's very much a "them" problem rather than one that we should fix by making the already rather lean "encourage folks to run" process more onerous. --Sohom (talk)20:53, 7 September 2025 (UTC)[reply]
3. It seems like ACE has trouble getting candidates, with a lot of the candidates signing up on the last day after people cajole them into it due to there not being enough candidates. In general, I think adding burdens/restrictions to processes that are having trouble is not a good idea. Careful of adding to the # of steps it takes to file a candidacy. –Novem Linguae(talk)03:55, 8 September 2025 (UTC)[reply]
4. After reading the other oppose votes I agree: making it easier to apply to be on the committee is much more important than reducing paperwork after the election is over. Better to delay it.Mrfoogles (talk)00:14, 9 September 2025 (UTC)[reply]
5. May dissuade late applicants. Also, there is no way for anyone to verify such claims. —xaosflux^Talk10:25, 9 September 2025 (UTC)[reply]

   Many (in some cases most) claims made by candidates in the nomination statement cannot be verified, and this ismore verifiable than the claim that someone is prepared to sign the agreement they are currently required to make.Thryduulf (talk)18:01, 9 September 2025 (UTC)[reply]

   OK, but it is verifiable by WMF staff - not something that the community can see. Now, if we wanted to make a requirement that candidates must complete the confidentiality process prior to the election, that could be verified - though would certainly prevent many late applicants. —xaosflux^Talk18:15, 9 September 2025 (UTC)[reply]

   Who has signed the agreements is verifiable by everybody once processed, which will be anybody who signs before a few days before the deadline (which can be fewer than the number who nominate themselves late in the period). For everybody else we will have to take it on trust - which is exactly what we currently do when candidates say they are prepared to sign (something that is verifiable by nobody). Additionally, it is exceedingly unlikely that someone will have correctly signed by the deadline but not have their nameon the list by the time voting starts, meaning that voters will have the choice not to elect someone who has not upheld their responsibility rather than only find out after they are elected. I'm genuinely struggling to see why this is not better in every way than the status quo?Thryduulf (talk)18:47, 9 September 2025 (UTC)[reply]

   I have no issue with someone not completing the agreement should they fail to be elected - negating its necessity. You mention that processing of agreements only takes "a few days", if so there doesn't seem to be any reason it needs to be done in anticipation. I do think it's a good idea for anyone considering running to do this well in advance, and think we should recommend it. —xaosflux^Talk13:11, 10 September 2025 (UTC)[reply]

6. We've got at least two weeks, if not a month, after-the-fact to sort this out. Also it's more work for WMF and candidates to solve a problem that we never had, which I'm sure no one wants. :)Alpha Beta Delta Lambda (talk)18:07, 9 September 2025 (UTC)[reply]

   I makes absolutely no difference at all to the workloads of either candidates or the WMF, and indeed gives the WMFmore time to do their side of things. The only difference to candidates is that it gives them a clear deadline to do their part, which impacts only them if they miss it. At present if they miss the (in my day at least) unclear deadline then it impacts the entire Committee.Thryduulf (talk)18:36, 9 September 2025 (UTC)[reply]

7. Too much looking to solve problems that do not exist. My vote would change if the supporters pointed to cases where this was actually an issueCzarking0 (talk)02:25, 10 September 2025 (UTC)[reply]
8. There is enough time for this after the results, and it doesn't seem to have been an issue in the past. Not worth adding an extra hoop for everyone to jump through at nomination time, given the difficulty getting candidates.the wub"?!"23:28, 10 September 2025 (UTC)[reply]
9. As others have noted, this feels close to a solution in search of a problem: there is no indication that the processing of this requirement has ever led to any significant issues in the post-election/pre-tenure period, and we don't need extra bureaucratic hurdles discouraging applicants. To be fair, the extra requirement would be very marginal to you average candidate for ArbCom, but then the issues with the current set-up are even more speculative and minor. On the balance, I don't see the benefit of moving this task earlier into the process flow.SnowRise ^{let's rap}00:36, 17 September 2025 (UTC)[reply]

• On the one hand, I support this. There is no reason for a candidate not to get this done early (if they're the type who puts stuff off until the last minute, that's not a good omen for their being an arb). Moreover, it would be a mess if somebody gets elected andonly then discovers there's something in the ANPDP which they find onerous and balk at signing. Which brings me to my real issue which iselection monitors/scrutineers/managers are not required to verify that this document has been submitted; that responsibility lies with the person or persons who grant access to the newly elected arbitrators. I don't know with whom the responsibility for checking should lie, but it shouldn't wait until after the election is over. If somebody were to win a seat without signing and then the activating steward has to hold up the process while they chase down the signature, we'd be right back in the situation this is trying to avoid.RoySmith(talk)14:44, 7 September 2025 (UTC)[reply]

  The only people whocan check are those who are responsible for receiving and verifying the signatures. Everyone can check them:Access to nonpublic personal data policy/Noticeboard but while everybody listed there has signed the agreement, not everybody who isn't hasn't: it is not updated in real time and, being maintained by humans, it is possible that there may be omissions. What might work is for someone to enquire atm:Talk:Access to nonpublic personal data policy/Noticeboard about any candidates whose names are not on the list a few days after they have nominated (archives suggest a typical response time of circa 3 days on that talk page).Thryduulf (talk)15:17, 7 September 2025 (UTC)[reply]

  To some extent, whatThryduulf said. More importantly, it is the expectation that a steward or a mailing list administrator (or in the case of Arbwiki, a wiki administrator) will in fact verify that any person receiving a (new) level of access has completed the agreement.Risker (talk)15:30, 7 September 2025 (UTC)[reply]

  if they're the type who puts stuff off until the last minute, that's not a good omen for their being an arb is a noble ideal, but one that conflicts with the age-old tradition of many or even most candidates signing up at the last possible moment.Toadspike[Talk]13:47, 9 September 2025 (UTC)[reply]

• Note that in accordance withfoundation:Legal:Wikimedia Foundation Confidentiality Agreement for Nonpublic Information/How to sign, I believe there are two confidentiality agreements that have to be signed? Perhaps one of the current arbitrators can confirm that they signed both theGeneral confidentiality agreement for nonpublic information (for non-VRTS related access to nonpublic information) andVRT users confidentiality agreement for nonpublic information agreements.isaacl (talk)17:32, 7 September 2025 (UTC)[reply]

  The critical one is the non-public confidentiality agreement. The VRT one can easily be signed after the election, as new arb orientation is not dependent on VRT access.Risker (talk)21:01, 7 September 2025 (UTC)[reply]

• Noting that, historically, new arb orientation has been delayed by up to 10 days in the past while awaiting this documentation completion. This is a big deal, when we are talking about a maximum 25-day window before the new arbitrator term begins, during which we also have year-end, ends of terms for educational organizations, multiple holidays, and sitting arbitrators endeavouring to clear the slate of existing matters, on top of carrying out new arb orientation. There is more to that orientation and onboarding period than meets the eye. I'm not sure I really want a candidate who can't be bothered to do the paperwork up front, although perhaps that is the perspective of someone who has had to work too often on teams and committees with people who didn't fully understand the assignment until it was too late.Risker (talk)21:11, 7 September 2025 (UTC)[reply]

  This has not been the blocker in recent years. If anything, the most recent timely issue of late has been two years ago when it took one of the stewards multiple weeks to complete scrutineering. If there's anything you should be voting for or against on the issue of timeliness, it's in fact proposal 2.Izno (talk)21:19, 7 September 2025 (UTC)[reply]

  Quite agree with you,Izno. Just haven't got that far.Risker (talk)21:26, 7 September 2025 (UTC)[reply]

• My tired brain isn't parsing this well (sorry if this is obvious to others!) - just to clarify: is this making a it a requirement for candidates to state they've done their side and are thus waiting for WMF to do their side, or for candidates to state they are already listed on the ID noticeboard? If it's the former, then I don't really see what impact this has on people throwing their hat in the ring at the close of nominations? WMF can process the submissions once nominations are closed; long before the results are announced. Am I missing something here?stwalkerster (talk)17:02, 9 September 2025 (UTC)[reply]

  The requirement is to have completed the three numbered steps atFoundation:Legal:Wikimedia Foundation Confidentiality Agreement for Nonpublic Information/How to sign. If you have already signed the agreement, then all you need to do is state you have done this in your nomination statement, a matter of a few seconds. If you haven't then it does add a time burden, but how much of one principally depends on how long you spend reading the actualpolicy but it's not a long document and anyone not already familiar with at least the basics of it probably shouldn't be applying for ArbCom at the last minute (which the guidance for candidates already discourages anyway). Setting up Phabricator account if you don't already have one may take 1-2 minutes (at absolute most), actually signing should take less than a minute. All told I'd be surprised if it took anybody longer than 30 minutes from start to finish, and most people significantly less than that, but I don't have any actual data to back that up. It's not impossible that Phabricator being down could prevent someone signing, but that's something the election commission are explicitly empowered to deal with on a case by case basis (the "Timeline" bullet point atWP:ACERULES).Thryduulf (talk)17:57, 9 September 2025 (UTC)[reply]

  Thinking about it a bit more, we already require candidates to state they are prepared to sign the policy when they nominate. This means that they must already have read and understood the policy to their satisfaction, so all that is needed is to quickly verify that the policy hasn't changed since they read it (which takes less than a minute), meaning the additional burden on candidates is even lower than I describe above.Thryduulf (talk)18:51, 9 September 2025 (UTC)[reply]

Proposal #4: Scrutineers to report sockpuppetry

Scrutineers are required to report any suspected sockpuppetry identified during the process of scrutineering the votes to two or more local checkusers who are not (a) a scrutineer (b) a steward, or (c) an arbitrator. Scrutineers may include available technical evidence in this report.Risker (talk)04:58, 10 September 2025 (UTC)[reply]

1. Striking votes of socks is part of the role of the scrutineers. However, sockpuppetry investigations based on technical evidence falls within the purview of local checkusers, who may determine appropriate next steps up to and including blocking of accounts. This is especially critical if the scrutineers are non-local functionaries, who do not have the ability to act on this information except with respect to striking a vote. Applying this rule also has the potential of reducing or eliminating the number of socks participating in the election.Risker (talk)04:58, 10 September 2025 (UTC)[reply]
2. * Pppery *_{it has begun...}04:59, 10 September 2025 (UTC)[reply]

1. I appreciate the sentiment here, but I think the stewards are a pretty clueful bunch and can figure out how to deal with the problems they find. If we want tosuggest that they turn over their findings to a local CU for further investigation, that's fine, but my guess is they already know that. The "required" language and the detailed specification of to whom they should report to is a step too far in my opinion.RoySmith(talk)12:19, 10 September 2025 (UTC)[reply]
2. I don't think the components of this are appropriate. Requiring scrutineers to individually pick 2 checkusers for private communications, and prohibiting them from communicating with arbcom or a non-arbcom local checkuser that happens to be also be a steward is too complicated. Actual socking in voting is publicly identified by way of voters being struck. I'd be OK with the local checkuser team being specifically authorized to perform a SPI on anyone that is struck. As far as odd edge cases where suspected socking is detected, which doesn't result in a strike - I'd be fine with empowering scrutineers to refer such information privately to arbcom as a group. I object to placing new requirements on scrutineers that they must do such referrals. —xaosflux^Talk13:05, 10 September 2025 (UTC)[reply]
3. I agree with Xaosflux.QuicoleJR (talk)19:00, 11 September 2025 (UTC)[reply]
4. Agree with Xaosflux.DreamyJazz^{talk to me |my contributions}08:18, 13 September 2025 (UTC)[reply]
5. per Roysmith, Xaosflux and my comments on proposal 5.Thryduulf (talk)10:32, 15 September 2025 (UTC)[reply]
6. per the above. And it's not like they don't report sockpuppetry; seeWikipedia:Arbitration Committee/Noticeboard/Archive 12#Edgar181 desysopped, for example.Some1 (talk)11:07, 15 September 2025 (UTC)[reply]

• Does this not already happen? If we do need a formal rule about it, and I'm not immediately convinced we do, then we should probably require the receiving CU to not be a candidate either.Thryduulf (talk)10:56, 10 September 2025 (UTC)[reply]

  I worked cooperatively with the scrutineers about specific scrutineering results last year. I think they sought me out because I was a CU and on ElectCom. Best,Barkeep49 (talk)20:58, 10 September 2025 (UTC)[reply]

• Aren't the scrutineers already required to report sockpuppetry, at least to ArbCom? Isn't that how they discovered the Edgar181 sockpuppets?Wikipedia:Arbitration Committee/Noticeboard/Archive 12#Edgar181 desysopped.Some1 (talk)10:57, 10 September 2025 (UTC)[reply]

  And I'd rather scrutineers report suspected sockpuppetry straight to ArbCom and not to two or more cherry-picked local checkusers, especially if the sockpuppetry involves admins.Some1 (talk)11:04, 10 September 2025 (UTC)[reply]

  Oh, this is just too funny. Up above, we have support for scrutineers to be from outside the project, which is defined previously as a "functionary" from outside the project. That does not mean it will be a steward. Up above, we also have strong support for keeping Arbcom's fingers out of any aspect of this election. But it's okay to report socks to them? Folks, please be consistent. If Arbcom is too conflicted to have any say about the election (to the point that people who they "supervise" shouldn't be scrutineers), then why would they be the people to which socking should be reported? And no, they aren't required to report sockpuppetry (and in fact, without an explicit requirement, it is easier to interpret the rules in a way that they *shouldn't* report it due to the supposed secrecy of all of the voting information). It seems that there is a real problem with consistency in the values being promulgated in this rule set.Risker (talk)13:44, 10 September 2025 (UTC)[reply]

  I don't see any inconsistency. CUs or ArbCom blocking socks has no impact on votes already cast, which must still be struck by scrutineers. I'm also confused byApplying this rule also has the potential of reducing or eliminating the number of socks participating in the election. in your !vote above – since scrutineerung happensafter voting closes, I don't see how any CU action at that point could reduce the number of socks participating in the election.Toadspike[Talk]21:26, 14 September 2025 (UTC)[reply]

  There are plenty of people who have multiple accounts (sometimes legitimate, sometimes not) that are eligible to vote. Just having a vote struck isn't as harmful to sock accounts as is having it formally reported. I'm fine with it being reported publicly (e.g., SPI) but that increases the workload of scrutineers (many of whom have easy access to CUs via email, IRC or Discord). Heck, make it one CU. I just want to make sure there is an authorized channel for scrutineers to report the socks they find outside of emailing Arbcom (which as a body is not expected to take any action on socking reports, and is officially only the oversight body for CUs).Risker (talk)01:06, 15 September 2025 (UTC)[reply]

  I would support this proposal if only one local CU was required, and if Arbs and scrutineers were not excluded.Toadspike[Talk]05:57, 15 September 2025 (UTC)[reply]

  On second thought, stewards who are local enwiki CUs are allowed to take CU action on enwiki, so should also not be excluded.Toadspike[Talk]06:02, 15 September 2025 (UTC)[reply]

Proposal #5: Scrutineers to report sockpuppetry (simplified)

Scrutineers must report any suspected sockpuppetry identified in the scrutineering process to an English Wikipedia CheckUser able to take follow-up action.Toadspike[Talk]06:19, 15 September 2025 (UTC)[reply]

1. Per my comments on proposal 5.Toadspike[Talk]06:19, 15 September 2025 (UTC)[reply]

1. Oppose requiring this. What do you expect the penalty for violation to be? Actual socking is already reported publicly via vote striking. I'm not so concerned aboutallowing such reports. —xaosflux^Talk10:23, 15 September 2025 (UTC)[reply]
2. If this was a proposal to add guidance for scrutineers saying something like "sockpuppetry should be reported to one or more English Wikipedia checkusers" I'd question the necessity (given this already happens) but wouldn't oppose. A formal policy requirement to report in a specific manner is too prescriptive imo.Thryduulf (talk)10:30, 15 September 2025 (UTC)[reply]

   @Thryduulf What do you mean by "in a specific manner"? I did not specify a manner, just that it must be reported. Also, based on Barkeep's comment above, this seems to have happened in the past solely on the initiative of scrutineers, which is not something I'd like to rely on.Toadspike[Talk]10:48, 15 September 2025 (UTC)[reply]

   The specific manner is "to an English Wikipedia CheckUser...", implying it must be sent to an individual who they know will be able to take follow-up action rather than, e.g. an SPI post or contacting ArbCom, or posting on the global CU mailing list, etc.Thryduulf (talk)11:01, 15 September 2025 (UTC)[reply]

   I read all of those as counting as "an English Wikipedia CheckUser" (or several). The wording is deliberately vague to make this requirement easy to fulfill.Toadspike[Talk]11:48, 15 September 2025 (UTC)[reply]

3. Oppose for much the same reason I opposed #4: We hire scrutineers because they are some of the most trusted people on the project. Let them do their jobs and don't try to micro-manage how they do it.RoySmith(talk)10:54, 15 September 2025 (UTC)[reply]
4. Agree with RoySmith.Some1 (talk)11:04, 15 September 2025 (UTC)[reply]
5. I think the proposal duplicates what anyone does if they suspect sockpuppetry, and what scrutineers already do when sockpuppetry is confirmed. --Minoa(talk)14:11, 16 September 2025 (UTC)[reply]