Wikipedia:Village pump (policy)

I would like to propose that all articles related to ethnicity in Afghanistan receive permanent extended confirmed protection restriction for one year.

This includes:

• all 34 provinces of Afghanistan + their capitals
• all 421 districts of Afghanistan + their capitals (if there is one)
• all of the cities mentioned in this list (if not already mentioned as a capital above)
• all of the ethnic categories mentioned in the article "Ethnicity in Afghanistan": Pashtuns (Pashtun tribes, Theories of Pashtun origin, and List of Pashtuns), Tajiks, Farsiwans (List of Tajik people), Hazaras (List of Hazara tribes, Aimaq Hazara, and List of Hazara people), Uzbeks (List of Uzbeks), Afghan Turkmens (Turkmen, Turkmen tribes, and Afghan Turkestan), Baloch of Afghanistan (Baloch people and List of Baloch tribes), Aimaq people, Nuristanis (Kalash people), History of Arabs in Afghanistan, Afghan Qizilbash, Pashayi people, Sayyid, Kyrgyz people, Gurjar (Muslim Gujjars and List of Gurjars), Pamiris (Wakhi people, Shughni people and Yidgha-Munji people), Brahui people, Afghan Tatars, Kazakh people, Parachi, Ormuri, Moghols, Jat people in South Asia, Peripatetic groups of Afghanistan, Hinduism in Afghanistan, Sikhism in Afghanistan, Jews in Afghanistan
• all historic provinces of Afghanistan mentioned here
• and articles on key ethnic and military-political figures such as all Naderi-Sayeds of Kayan, Abdul Rashid Dostum, Haibatullah Akhundzada, Ahmad Shah Massoud, Burhanuddin Rabbani, Abdul Karim Brahui and Abdul Ali Mazari.

This is a continuation of this discussion.

I frequently see editors accuse others of SYNTH in protected situations, like synthesis in exercising editorial judgement or in creating a collection of RS and analyzing them for ourselves, of violating WP:SYNTH. On Wikipedia:What SYNTH is not, it clarifies SYNTH is not literally the word synthesis, but I want to add some clarifications to it based on some great suggestions from WhatamIdoing here.

WP:ONUS says:

The responsibility for achieving consensus for inclusion is on editors seeking to include disputed content.

WP:NOCON says:

When discussions of proposals to add, modify, or remove material in articles end without consensus, the common result is to retain the version of the article as it was prior to the proposal or bold edit. (emphasis mine).

Which is it? There's currently an edit war over inclusion of some contentious material, which has been in the article for some time. At this point, there has not been agreement on whether to keep it, meanwhile several editors are alternately removing and restoring it as per Wikipedia policy, citing the above reasons.

The original discussion went on for a few months this fall, without reaching consensus, and was dormant for most of the winter, so it seems reasonable to say that the discussion "ended without consensus", even though it is still ongoing.

So, do we keep or remove the disputed material pending discussion? And if it ends without consensus do we restore/keep it as per WP:NOCON?

Looking for policy decision independent of the details of the dispute. BTW, it's not one of the exceptions in WP:NOCON. Mr. Swordfish (talk) 13:06, 24 April 2026 (UTC)

I have no idea what the dispute is that you're referring to, but in general I'd think ONUS trumps NOCON in this regard. Partly because it's very explicit on the point, while NOCON is a broader more general policy. And also because the wording of NOCON is not expressed as a requirement, merely that "the common result is..." In general, the project does not particularly strive for protection of status quos, rather we seek to preserve verifiability and NPOV, erring on the side of caution where things aren't clear or people can't agree.  — Amakuru (talk) 13:14, 24 April 2026 (UTC)

They're not mutually exclusive. NOCON says that the common result is to retain the prior version, not that it must do that. If there's an ONUS challenge then ONUS takes priority because there is no consensus to include.

All you have to remember is that Wikipedia's default state is "consensus to include" and therefore anything that doesn't have consensus can and should be removed. Rambling Rambler (talk) 13:17, 24 April 2026 (UTC)

We have discussed this MULTIPLE times over the last few months… and unfortunately have not (yet) reached a consensus. Please look at these discussions on the archives.

If I had to summarize where we currently stand, we invoke WP:WRONG VERSION. Meaning that, after the initial flurry of removal, reverts and re-reverts… we STOP EDIT WARRING and freeze the article in one state or the other while we discuss. Sometimes this means the disputed material remains in the article while we discuss (favoring NOCON) and at other times we omit it during the discussion (favoring ONUS). Which we do often depends on why the material is disputed.

Most editors (not actively involved it the dispute) don’t really care whether the material is left in or taken out during discussion… what matters in the long run is that the discussion takes place and reach we a consensus on what to do after the discussion ends. Blueboar (talk) 13:37, 24 April 2026 (UTC)

I also have no idea what the background for this dispute is. but my reading of the policies is that the general rule is WP:NOCON while WP:ONUS is an exception for a specific case, namely disputes related to verifiability. So where WP:ONUS applies it follows that it overrules WP:NOCON. Similar situation as with WP:BLPUNDEL saying that if the removal of content is related to a BLP then the burden is on whoever wants to include it to ensure that it complies with content policy, and that If it is to be restored without significant change, consensus must be obtained first. So, if there is a dispute over verifiability (a reasonable one, you can't claim anything you feel like is a verifiability dispute obviously) then WP:ONUS applies, otherwise WP:NOCON applies (with the exception given by the policy for discussions on BLP material, copyright matters and external links).
⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 14:06, 24 April 2026 (UTC)

No… BURDEN relates to cases involved with Verifiability. ONUS relates to other disputes. Blueboar (talk) 19:09, 24 April 2026 (UTC)

In my view it makes no sense to consider WP:ONUS to apply to every no consensus situation except for that of verifiability when it's a part of WP:VERIFIABILITY rather than WP:CONSENSUS, and WP:CONSENSUS has a section, WP:NOCON, that is specifically meant as a broad rule for no consensus situations.

I can see an exception to this in cases where there was never any consensus to include a piece of content in the first place, as that would just be another case of a broad rule in some situations being overruled by a more specific one (although I find even this a bit dubious as this would also lead to it making no sense that WP:ONUS is part of WP:VERIFIABILITY rather than WP:CONSENSUS), but if a consensus to include something has been established at some point, and later there is no consensus among editors on whether or not to include it, then WP:NOCON has to take precedence over WP:ONUS. I suppose these policies could be better worded for clarity (as could many other of our PAGs), but I don't think they're necessarily contradictory as it stands. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 03:36, 25 April 2026 (UTC)

I agree that having ONUS in the WP:Verifiability policy page is confusing - this is why I (and others) have argued that it should either be moved to a different page (such as WP:Consensus) or deprecated completely.

However, if you look at the history of when and how it was adopted (and how most editors have interpreted it since then) it becomes a bit clearer that WP:BURDEN is for Verifiability disputes while ONUS is talking about the need for consensus in disputes other than Verifiability. This is why they are placed in two distinct sections and have different shortcuts. Blueboar (talk) 12:44, 25 April 2026 (UTC)

Burden is specific to whether something is verifiable, but says nothing about where or how or when verifiable material belongs n any specific article. Alanscottwalker (talk) 13:00, 25 April 2026 (UTC)

I hear what you're saying but to me this is just a matter of there not being much of an option. It's not just confusing, it results in a farcical situation. If I am to interpret WP:ONUS as applying to everything but verifiability, then how can WP:NOCON, that is specifically intended as a broad rule for no consensus situations, which says When discussions of proposals to add, modify, or remove material in articles (the bolding is part of the policy, not added by me) be interpreted in a way that doesn't make a total joke out of our policies? It very, very clearly says remove material in articles, and I don't see how that wouldn't be a direct contradiction of WP:ONUS if I accept your interpretation. Between interpreting it as two of the most important policy pages we have, which proudly state that they have been thoroughly vetted by the community, being outright irreconcilably contradictory, and interpreting it as WP:ONUS only applying for verifiability and/or only in cases where there was never any previous consensus for conclusion established, I don't see much choice but to go for the latter. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 14:03, 25 April 2026 (UTC)

You know what just nuke WP:ONUS and call it a day. All for? No? Worth a try. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 14:06, 25 April 2026 (UTC)

It's been tried. WhatamIdoing (talk) 21:35, 25 April 2026 (UTC)

If at first you don't succeed.. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 04:02, 26 April 2026 (UTC)

• Loathesome though it is to try to answer these matters without a context, I would say that in principle, imagining some hypothetical case where no other considerations apply, then the core content policy prevails in articles, and the consensus evaluating policy prevails in policy decisions. So mainspace, WP:ONUS, and Wikipedia space, WP:NOCON.—S Marshall T/C 16:18, 24 April 2026 (UTC)

I intentionally didn't include the context since I was looking for an answer independent of this particular example, but since several have said that context matters, the issue is at List_of_topics_characterized_as_pseudoscience. See recent history and the talk page for details. Mr. Swordfish (talk) 18:20, 24 April 2026 (UTC)

Thanks for the note. Context matters in two ways: One is that NOCON has a couple of specific exceptions. The other is that in my (not inconsiderable) experience with this specific WP:PGCONFLICT, editors like to see the discussion so they can evaluate the discussion, in the hope that they can declare that there's no conflict because there's "obviously" a consensus for ____. WhatamIdoing (talk) 21:37, 25 April 2026 (UTC)

• I see no conflict here at all. If there is disputed material, then the "bold edit" being contested would be the edit that added the disputed material. To retain the version of the article as it was prior to the proposal or bold edit, per WP:NOCON, would be to retain the version without the disputed material until there is consensus for its inclusion, just as WP:ONUS requires. BD2412 T 18:47, 24 April 2026 (UTC)

  I'm not sure this matters, but the edit that added the material was almost 5 years ago. Searching the archives, it wasn't up for discussion on talk page except one brief mention two years ago. Then there was a discussion last fall that failed to reach consensus to remove the material.

  So, my take is that the "bold edit" was removing long-standing material rather than adding it. WP:NOCON addresses removing as well as adding material. If it only applies to adding material, then it should say so. Mr. Swordfish (talk) 18:59, 24 April 2026 (UTC)

  Yes, it was. Nevertheless I'd keep it out, both on the policy grounds that I said above, and also on the facts. Historical materialism doesn't seem pseudoscientific to me. It's wrong, but I think the claims it's making aren't really masquerading as science and aren't easily confused with science.—S Marshall T/C 20:43, 24 April 2026 (UTC)

  Example of the "Core content policies prevail on content pages" approach. Four years old but, I think, really clear.—S Marshall T/C 20:57, 24 April 2026 (UTC)

I think it important to point put that nothing in ONUS states you can remove content, you still need a good reason to dispute the content. My answer to the general question is that Wikipedia doesn't have a deadline and editors should try consensus building rather than edit warring over whether to include or exclude the content while that happens. -- LCU ActivelyDisinterested «@» °∆t° 22:57, 24 April 2026 (UTC)

Also make sure to check the talk page archives to ensure a prior consensus to include was reached, as that would satisfy ONUS. -- LCU ActivelyDisinterested «@» °∆t° 23:00, 24 April 2026 (UTC)

Also… check to see if there was a prior consensus to exclude/omit… as that would also satisfy ONUS (but on the opposite direction). Blueboar (talk) 13:11, 25 April 2026 (UTC)

Moreover, ONUS certainly trumps NOCON, but only in cases where the proposition "this content is unverified" is actually agreed upon. If there isn't some form of rough consensus for that, then this seems likely to lead to all sorts of unpleasant gamesmanship. Sławomir Biały (talk) 13:51, 25 April 2026 (UTC)

If the content is uncited, then WP:BURDEN (not ONUS) applies, and the content can be removed until it is WP:Glossary#cited.

If the content is cited but there is a consensus that the citation {{failed verification}}, then the citation can be removed, and editors (←note intentional use of plural) should look for a new source.

If no reliable source can be found, then the material is WP:Glossary#unverifiable (to the best of our knowledge) and should be removed per WP:V (not ONUS). WhatamIdoing (talk) 21:41, 25 April 2026 (UTC)

Although some uses of ONUS cause controversy, this is no different from any other part of policy. Whether it's bad faith blanking stating ONUS or stonewalling using QUO/NO ON. ONUS may not be universally accepted but neither are other parts of policy that are used by the majority of editors without issue. -- LCU ActivelyDisinterested «@» °∆t° 19:32, 26 April 2026 (UTC)

Delete ONUS in its current form

At the risk of a tangent, ONUS should either be limited to when verifiability is the reason for removing/challenging content, or it should be moved into NPOV or EP. When people want to say "the onus is on you to resolve the verifiability/citation issue" they cite WP:BURDEN. ONUS is kind of only cited when verifiability isn't an issue. — Rhododendrites ^talk \\ 13:07, 25 April 2026 (UTC)

I don't think that's quite accurate, you would also cite onus, where adding a verifiable bit of information creates a fictional narrative, or in answering the issue of where verifiable bits of information belong together to create a varfiable whole. Alanscottwalker (talk) 13:45, 25 April 2026 (UTC)

Sigh… we have discussed moving ONUS out of WP:V before… The problem is that we can not reach any consensus on where to move it to. Blueboar (talk) 14:11, 25 April 2026 (UTC)

Let's move it to what is now the redirect WP:ONUS and then add this template at the top of the page. That ought to do it. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 14:40, 25 April 2026 (UTC)

The general notion behind ONUS is that the person pushing bad content, especially of the pseudoscientific variety, has to do the work of proving that the community wants their content in the article.

This is the goal:

• Hal Homeopath: Homeopathy cures everything, even heroes who are mostly dead!^[1]
• Dave Defender: Revert nonsense.
• Hal: Prove to me that the community doesn't want this valuable information in Wikipedia.
• Dave: Nope. It's your job to prove that the community wants this, and until you prove a positive consensus for inclusion, that nonsense stays out.

WhatamIdoing (talk) 21:47, 25 April 2026 (UTC)

Well, if that's the case, would you support moving it to WP:FRINGE and modifying it appropriately to reflect that? --Aquillion (talk) 04:58, 26 April 2026 (UTC)

At this point, after years of harping on this problem, I would support any change that (a) resolved the contradiction and (b) the community was actually willing to stand by. WhatamIdoing (talk) 17:51, 29 April 2026 (UTC)

Not all bad content is FRINGE content though. Some might just be based on primary sourced historical accounts, like someone saying teh Britons descend from Brutus of Troy (or something similar but less obviously wrong). It is not a policy just for fringe content. Sirfurboy🏄 (talk) 18:16, 29 April 2026 (UTC)

The interpretation I was just arguing against was that WP:ONUS doesn't apply to verifiability disputes, which are covered by WP:BURDEN, while WP:ONUS applies to every non-verifiability dispute. What you've described seems to pretty clearly be a verifiability issue. So while the interpretation of WP:ONUS I oppose does nothing to prevent the situation you described, my interpretation of WP:ONUS actually would, and it would do so without also meaning that every piece of content without explicitly established consensus can be summarily removed with for nearly any reason at all, forcing editors to get consensus before restoring the content. Either way, let's forget about interpreting what is and look at options for what can be, because the fact this discussion is taking place is more than enough to tell us that the policies are not clear enough. A new editor should be able to read and understand them; we're talking about two of the most important policies we have here, WP:V and WP:CON. Some ways we that prevent situations like those you described without giving a carte blanche to removal of longstanding content:

• Deprecate WP:ONUS and make WP:QUO a part of WP:NOCON, potentially with some wording change (it's already referenced by it).
• Deprecate WP:ONUS and add a line to WP:NOCON saying something like If a dispute is over addition of content to an article then the inclusion of that content is considered to have no consensus even before the end of discussion.
• Move WP:ONUS to WP:CONSENSUS under WP:NOCON and change it to say If the addition of content to an article is challenged the person wishing to add said content must obtain a consensus for it.

The last two should have the same effect, unless I've somehow missed something. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 07:05, 26 April 2026 (UTC)

I once closed a RFC about an incredibly cute creature called the bee hummingbird. This adorable miniature bird is the world's smallest bird. It's also, according to reliable sources, the world's smallest dinosaur.

This fact was inserted into the lead of bee hummingbird and the inserter wouldn't allow it to be removed without a full RFC.

ONUS is the rule that lets us remove verifiable material from inappropriate places. I think that's why it belongs in WP:V. WP:V is where people would naturally go to check whether they could remove verifiable information that's in the wrong place.—S Marshall T/C 09:03, 26 April 2026 (UTC)

Alright, then how about changing ONUS to say If an editor adds content to an article and that addition is challenged by another editor, the editor that wants to include the content has the responsibility to get consensus for inclusion. and keeping it at WP:V and inserting a hatnote at the top saying see also WP:NOCON? ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 10:22, 26 April 2026 (UTC)

That could be read two ways.

1. If you add material that's verifiable but in the wrong place, then I get to remove it from the article, and it stays out until you get consensus otherwise.

2. If I find information that's verifiable but in the wrong place, then I get to remove it from the article even if it's been there for a long time, and it stays out until someone gets consensus otherwise.

At the moment it means 2. Do you feel that it should be changed to mean 1?—S Marshall T/C 11:02, 26 April 2026 (UTC)

I think I understand what you mean when you say the policy means that you can remove something that's verifiable but in the wrong place, and yeah of course I want people to be able to do that when it's appropriate to do so, but my problem with it is how WP:ONUS is essentially allowing anybody to claim something is in the wrong place for any reason. It's a free ticket to force content to be kept out until consensus is established, as long as you have some sort of justification that isn't blatantly ridiculous, and those who know their way around don't have much trouble pulling that out of their hat.

Also the first of my suggestions would keep the ability to do 2 intact, but it does so while not forcing editors to keep articles in damaged states while trying to prove that there is consensus for inclusion. If their attempts end up resulting in a no consensus discussion then by all means, go ahead and apply WP:ONUS.

Anyways, if it turns out I'm in the minority here and people don't want to change it we have to address why WP:NOCON is contradicting it and fix that. Somethings gotta go! ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 12:27, 26 April 2026 (UTC)

Here's the thing: it actually depends who's wrong.

Let's say Anne finds a fact that she thinks should be widely known. It concerns, for example, the exact shade of hair dye used by Donald Trump. Anne adds this to the lead of Donald Trump. Charlotte reverts that edit, claiming it's not important enough to be in the lead. Policy ought to be on Charlotte's side.

Let's say Bob finds a fact that he thinks should be less prominently published on Wikipedia. Say, Bob doesn't think it's important that Donald Trump is the President of the United States, and wants to move it out of the lead of Donald Trump. Charlotte reverts that edit, claiming that Trump's presidency is important enough to be prominently stated in the lead. Policy ought to be on Charlotte's side.

In both cases, what makes Charlotte right isn't sources or facts. It's her editorial judgment about what belongs in the lead.

Tragically WP:EDITORIALJUDGMENT is still a redlink.

There's a perfect phrasing for WP:ONUS, but nobody knows what it is. I personally would oppose weakening it or downgrading it.—S Marshall T/C 13:54, 26 April 2026 (UTC)

Right, but my suggestion literally makes policy be on Charlotte's side in both cases. In the first one Anne will have to get consensus to add the content, because the status quo/stable version is the version without the content, and in the second one Bob will have to get consensus to remove the content, because the status quo/stable version is the version with the content.
I would think the current WP:ONUS policy would give Bob the right to do that if the example wasn't so blatantly incorrect by him. Perhaps he would look at the part of Trump's lead that currently says "In response to the COVID-19 pandemic in 2020, he downplayed its severity, contradicted health officials, and signed the CARES Act" and decide that Trump contradicting health officials was little more than a footnote of how he handled the pandemic, and unfit for the lead, with reference to WP:LEADREL. Maybe he even feels that downplaying the severity isn't that important in comparison to the signing of a law. After all, laws are what actually do stuff, right? Perhaps he's incorrect, but it's not so obvious as to say he is being blatantly disruptive, so when Bob invokes WP:ONUS Charlotte can't do much other than take it to the talk page and try to gain consensus, and until then the lead says "In response to the COVID-19 pandemic in 2020 he signed the CARES Act." (example might be wrong as there might be consensus somewhere on the talk page for writing the lead like this but you get the idea). ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 14:17, 26 April 2026 (UTC)

If it isn't obvious I'm insinuating that Bob might not have entirely pure intentions in removing parts of the lead that make Trump sound bad, but unfortunately for Charlotte WP:AGF applies. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 14:40, 26 April 2026 (UTC)

I think your proposal is a good one for articles that are already right. For articles that are wrong and/or have been wrong for a long time, it's less so.

Real cases about this tend to be horribly messy with many complicating factors, so I don't have examples that isolate this one issue without raising others. But, for example, I'd ask: how would you have closed this RfC?—S Marshall T/C 16:35, 26 April 2026 (UTC)

I will read it and get back to you on that. Certainly won't complain about the subject considering I'm currently studying economics. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 16:54, 26 April 2026 (UTC)

You're getting a long post here I'm afraid. Comes with the territory of asking me to analyze something. Anyways, the first part of that discussion _{(you say RfC, I'm not sure if you're asking me to think of it as an RfC or if it was an accident, because it's not an RfC from what I can tell).} is an absolute circus. I'm not even sure how many civility rules were broken and there's even someone socking and disrupting discussion there. In the midst of all of that shouting I did spot one point that sums up why this discussion is bound to be a mess: The point is that "rent control" is too broad and diverse a set of policies for this categorical statement to be meaningful. which is very true, and this whole dispute would have been resolved a much better way if people directed their focus to crafting a better version of the text. The whole issue with it is that it uses consensus about a generalization without explaining what situations economists view the generalized statement as applying to and which ones they view as exceptions. Also I'd have a hard time closing that thread instead of joining in as I see multiple flaws in some of the arguments made that nobody else has pointed out. If I had to close then I suppose I'd have to call it no consensus, and I'm assuming the reason you asked me to look at this in the first place is because you wonder what I think should happen as a result of the no consensus.

The claim was added to Rent control in the United States on 24 December 2020. It was challenged 6 days later by the addition of a citation needed tag. The addition of the tag was reverted, and then the editor who added it instead changed the wording on 6 January 2021 (the most notable thing to happen that day, I'm sure) which was reverted right away. Clearly a content dispute is in progress less than two weeks after the addition of the content so I'd say that it should be removed if there is no consensus to include it. It's clearly tricker with Rent regulation as it appears to have been around for quite a while on that article, being added 14 September 2019. I will note that the edit summary says Economists' views: lede must note the consensus among economists that rent control is crap which tells me that the person adding this content wasn't super familiar with the nuance economists tend to give this matter. It was challenged on 27 October 2019 and on 9 December 2019 via the addition of citation needed tags. It was thenremoved on 27 April 2020. To me there wasn't enough time nor enough editors leaving it alone to call that a de facto consensus to include it. Only a handful of editors made non-automated edits outside of sections where the claim wasn't located between the addition and the first removal, let alone the first challenge.

All in all I'd get rid of both claims until consensus can be established for including them as neither of them have much of a claim to having gained the community's silent acceptance for being included. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 21:29, 26 April 2026 (UTC)

I asked you to think about it because your answer tells me a lot about how you understand ONUS and NOCON. I'm intrigued to see that in practice, you seem more inclined to get rid of disputed claims than I am. (This is how I closed it later that day.)—S Marshall T/C 22:50, 26 April 2026 (UTC)

Oh, I guess I should have noted that if it wasn't a hypothetical and I had to go off the rules as the consensus is to interpret them I would have gotten rid of both claims per WP:ONUS after closing with no consensus, I suppose. The whole de facto consensus thing is my own idea and how I want policy. to be interpreted after all. I wouldn't try to unilaterally enforce that on the community.

Also I have to say that's a clever close. For me it was on the edge of no consensus and consensus for inclusion, and considering how the discussion was anything but straightforward, with a good deal of conduct violations, disruptions and some arguments than were pretty WP:OR-ish. In that context I'm more inclined to go no consensus than I otherwise would be. I just couldn't confidently extract a meaningful consensus there. I have to say it's highly commendable that you were able to. Goes to show the difference in experience between us I suppose (and if I may say so I think I can safely say most editors are thankful that we have experienced closers such as yourself willing and able to deal with messes like that discussion). ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 05:09, 27 April 2026 (UTC)

Thank you for those kind words. Does thinking about that real life example affect your view of WP:ONUS, at all?—S Marshall T/C 07:23, 27 April 2026 (UTC)

Okay uh, my bad for not responding to you sooner but I was doing some other stuff (it was also far from clear there was a reply to me here as the talk page got all wonky and I had to refactor it to make it make sense). Yes, my view on WP:ONUS has been affected by this discussion (ours and the topic as a whole). I do believe that the idea I had of what WP:ONUS should be does work out as I intended when I applied to that real life example, but it has made me think there are likely situations where it would not, and that those situations might have troublesome consequences, especially when covering two articles like that. Under different circumstances that article where it was longstanding could have been there long enough to have a de facto consensus under my idea, and that would lead to the claim remaining in one article but not the other, which is nonsense. Also it was hard to apply, and, considering that "always apply this policy the way Maltazarian would have" would unfortunately not work as a policy, that makes me unsure about how consistently it could be applied, and as this is not simply procedural a lack of consistency is also a lack of fairness.

That has made me rethink, as I still want to protect against longstanding content being removed and having to wait for consensus to be added back in. What I've come up with is that we continue removing content unless there is consensus for inclusion, but while the discussion over whether to include it or not is ongoing the content remains in the article if it is longstanding. That way any addition can be removed and won't be visible on the page while discussion is ongoing, and any longstanding content can't be removed and be forcefully excluded until a discussion finishes. In both situations, if no consensus for inclusion is achieved, the content is excluded. So WP:ONUS would remain largely unchanged except for a procedural caveat, and WP:NOCON would be changed to not be in conflict with WP:ONUS. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 15:22, 29 April 2026 (UTC)

Building on editorial judgement, it looks like part of the difficulty is that some disputes often aren't about verifiability, but about weight. In those cases, something can be sourced but not prominent across sources, which is where ONUS and NOCON pull in different directions. Real issue is due weight. When guidance isn't very explicit, we can end up going round and round the same point leading to outcomes that vary across similar discussions. I notice something similar around NPOV/due weight too and will start a separate discussion. NicoR8 (talk) 17:21, 28 April 2026 (UTC)

Maybe WP:ONUS should simply specify:

• Had the disputed content been on a page for over a week?
  • Yes (content has existed for over a week): WP:ONUS means the burden of proof is on the party seeking to remove content from the article to defend removing the content and the content is kept in the meantime with a disputed content, discuss here tag.
  • No (content was just added):
    • Is the page high visibility?
      • Yes: WP:ONUS means the burden of proof is on the party seeking to add content to the article to defend adding the content and the content is removed in the meantime.
      • No:
        • Is the content plausible/not outrageous?
          • Yes: WP:ONUS means the burden of proof is on the party seeking to add content to the article to defend adding the content and the content is added in the meantime with a disputed content, discuss here tag.
          • No: WP:ONUS means the burden of proof is on the party seeking to add content to the article to defend adding the content and the content is removed in the meantime.

Alexandraaaacs1989 (talk) 03:47, 29 April 2026 (UTC)

Something is out of order here. S Marshall's outdented reply does not appear to be to this. Your proposed change to ONUS is unworkable - indeed it is a repeal and suggestion to replace with WP:FAITACCOMPLI. In your first branch you are saying that if someone can sneak in an edit and no one notices for a week, the edit should now stand until someone can get a consensus to remove. Perhaps you only edit highly traffic articles, but I've removed nonsense from pages where the nonsense added was the only edit in the last nine months. Forcing me to start a talk discussion on a page almost no one is watching would be a ridiculous thing to do. It's a vandal's charter, and a gift to POV pushers everywhere. Just no.

But you go on. Even if the content is less than a week old, you suggest there is only an onus on those adding content if the page is "highly visible" (who judges what is highly visible) or if the content is not plausible and is outrageous (and how are we to judge that?) ONUS forces a discussion where there is already a reasonable challenge. ONUS cannot be invoked alone. ONUS may only be used where the information is being challenged in its own right. In such cases we have a policy (ONUS) that requires editors to discuss. It works. Don't mess with it. Sirfurboy🏄 (talk) 08:01, 29 April 2026 (UTC)

For the record: the part saying "something is out of order here" was in reference to a version of the talk page that has since been refactored to address this (refactoring diff). ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 14:58, 29 April 2026 (UTC)

I agree that it might be true that there is a perfect phrasing for WP:ONUS, but nobody knows what it is. That said, I think one option that would be preferred by FRINGE fighters would sound like this:

• The responsibility for achieving consensus for inclusion is on editors seeking to include disputed content. →
• If material is removed from an article for a plausible reason (e.g., for any common-sense or policy-based reason, but not unexplained removals, which could be section-blanking vandalism), then the editor who wants to include the disputed content has the duty to demonstrate a positive consensus for inclusion. Having articulated any plausible reason at all for removing the content, the editor who removed it has no further responsibility for determining its inclusion or exclusion.

WhatamIdoing (talk) 18:13, 29 April 2026 (UTC)

I'd be fine with this if we added a clause that said something like:

If the disputed content is longstanding it should be kept in the article while discussion on whether to include it or not takes place. This does not change the fact that responsibility for achieving consensus is on editors seeking to include disputed content.

It should replace "longstanding" with an actual period of time that the content has to have been included. This will obviously be somewhat arbitrary but it will be fought over otherwise. I think I'd draw the line somewhere in the range of 6 to 12 months. To be clear if a discussion ends in no consensus the content would still be removed, but at least this protects some key content from being removed and the article being forced into a damaged state for a prolonged period of time due to stonewalling by those against inclusion. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 18:35, 29 April 2026 (UTC)

Why would you want to keep old garbage in the page? Have you looked at Wikipedia:List of hoaxes on Wikipedia, and considered what "keep longstanding garbage until discussion finally agrees to remove it" would mean? WhatamIdoing (talk) 23:20, 29 April 2026 (UTC)

Hoaxes aren't verifiable. Kolya Butternut (talk) 10:05, 1 May 2026 (UTC)

Hoaxes aren't verifiable in reliable sources, but they can be cited, which means that BURDEN (which applies exclusively to uncited information) wouldn't be enough, and to remove content for verifiability alone would require some sort of consensus that it actually is unverifiable. WhatamIdoing (talk) 05:09, 3 May 2026 (UTC)

If anything, I think this would be better:

If the disputed content is longstanding, it should be kept in the article while discussion on whether to include it takes place. This does not change the fact that the burden of proof is on editors seeking to include disputed content. However, the editor seeking to remove the content has the responsibility of starting the discussion and notifying other talk pages of the discussion as needed. In the meantime, a tag should be placed on the article next to the disputed content, indicating the content is being discussed on the talk page.

That way, the person contesting consensus has to put in some effort getting the discussion started. I think it's much fairer that way to both parties involved. There's also the important point that a lot of the time when info is simply deleted from a page, people might not notice. So making the person deleting content have to notify the talk page ensures what would otherwise be a deletion discussion might fly under the radar unnoticed. Alexandraaaacs1989 (talk) 23:50, 1 May 2026 (UTC)

• Lets break this down:

Editors are allowed to add verifiable information to articles. They do not need to gain prior consensus to do so.

Other editors are allowed to remove verifiable information from articles. They do not need to gain prior consensus to do so.

Sometimes this results in a dispute. At that point, editors discuss and try to reach consensus on whether to keep or omit (also considering other options such as rewrite, move it to another article, etc). If that discussion results in a consensus, we do whatever that consensus says.

The problem arrises when the discussion results in “No Consensus”. One side will argue “there is no consensus to keep” while the other side will argue “No, there is no consensus to remove”. Wikipedia has never been able to come up with clear “rules” on what we should to do in “No Consensus” situations, but ONUS is the closest we have come to doing so.

However, ONUS is far from universally accepted… we have had numerous discussions on whether to keep it, remove it, change it, move it, etc… And (ironically) the result of all these discussions is No Consensus.

Which leads to an interesting catch 22… ONUS says we need consensus to keep… yet we have no consensus to keep ONUS… so (technically) ONUS should be removed from the policy… per ONUS! Blueboar (talk) 14:11, 26 April 2026 (UTC)

That's actually a pretty funny point I hadn't even though of. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 14:18, 26 April 2026 (UTC)

The Wikilawyer response to this is: ONUS (as part of WP:V) only applies to article space, not policy space. But I thought it was amusing too. Blueboar (talk) 14:54, 26 April 2026 (UTC)

Yes I did immediately think of that but I didn't want to ruin it. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 15:20, 26 April 2026 (UTC)

Your two first paragraphs are incorrect, because 1) verifiability is not the sole requirement; and 2) by editing, they are necessarily asserting a consensus that their editing in good faith meets the requirements of the pedia. -- Alanscottwalker (talk) 14:59, 26 April 2026 (UTC)

That isn’t how it works for the average editor. They just add stuff and remove stuff. They don’t need permission from the community to do either. It’s only when that basic freewheeling system breaks down and someone objects to an edit (addition or removal) that we worry about the “requirements of the pedia”. People want reasons for the objection after all. Blueboar (talk) 16:25, 26 April 2026 (UTC)

Thus, an edit is an assertion that it complies, because it is immediately a part of the topic for all the world to see. Others want good reasons for the edit too: the edit often explains itself, and a note can be appended to the edit. When others don't understand why or how the edit complies with requirements, it is more likely to raise objection. Alanscottwalker (talk) 10:34, 27 April 2026 (UTC)

But any old editor can come along and object to any information they don't like on Wikipedia and delete it pointing to WP:ONUS, forcing others to start an arduous consensus-building process if they want to re-add it, even when that point was already supported by presumed community consensus for years. The person seeking to upend the status quo's consensus (which has existed for X amount of time, whether a week or otherwise) should start a talk page discussion about whether to remove the material, and consensus from that discussion should determine whether to update community consensus not to include that material. In the meantime, disputed material should have an appropriate tag labeling it as such. If you disagree, I'd like to hear a better solution to drive-by ONUS deleters. I am all ears. Alexandraaaacs1989 (talk) 10:03, 29 April 2026 (UTC)

Is it better to change the rules so that if any old editor can sneak their preferred POV into an article and keep it for a week, then it becomes unremovable without spending quite silly amounts of time and effort?—S Marshall T/C 10:12, 29 April 2026 (UTC)

That's the purpose of a "disputed content" tag we'd add to the disputed content - to recruit attention to the content dispute on the page. And it's pretty rare editors will make egregious POV pushes in high-visibility articles without immediately getting detected and corrected. And I still haven't heard a response to the drive-by ONUS deletion concern. Alexandraaaacs1989 (talk) 15:16, 29 April 2026 (UTC)

There isn't a satisfactory response to the drive-by ONUS content removal concern. (I'm sure you know that "deletion" has a technical meaning, so for clarity, I say "content removal" for something that any editor could simply revert.)

In policy terms we have to make a decision. Either option has serious drawbacks. We have to think about how this could be used to advance a position in fields like US politics, where we have to deal with a lot of people who have agendas.

We definitely need to remove disputed content while we discuss it, if it concerns or relates to living people or groups. That's BLP and not negotiable. We really ought to do the same in CTOPs. But what the rule should be outside those high stress areas is something on which reasonable people could differ.—S Marshall T/C 15:27, 29 April 2026 (UTC)

Just to make sure you're not missing it due to the reformatting I did: I replied above in that discussion we had about the example you gave a few days ago with a new idea that keeps the responsibility on those wishing to include content but while also giving a layer of protection to longstanding content. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 18:55, 29 April 2026 (UTC)

On the "disputed content" tag, my experience is that the same people who will put disputed content back into an article will often also remove any such maintenance templates. ONUS forces the creation of a consensus to include, where a dispute exists. It is not broken. Sirfurboy🏄 (talk) 16:15, 29 April 2026 (UTC)

But why is consensus not also required when excluding/removing long-standing relevant content? ~2026-24524-49 (talk) 19:06, 29 April 2026 (UTC)

The thing about "relevant content" is consensus is needed to determine what is and isn't relevant. The main concern with putting the responsibility of consensus on those wanting to remove content is that it gives a lot of power to people sneaking in WP:FRINGE and WP:UNDUE content into articles. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 19:22, 29 April 2026 (UTC)

Consensus is required to include, even for longstandng content. Those who think it isn't are in error. But, in practical effect the difference may be small. Much longstanding content already has a clear consensus, either explicit or implicit. It may have been discussed with explicit consensus, or it may have been worked on by multiple editors, in which case there is a clear implicit consensus. Onus requires a consensus to be reached on inclusion, but once there is a consensus, it cannot be over-ridden without a consensus at an equal or greater WP:CONLEVEL.

In practice, if text is longstanding in an article and disputed, an RfC may be needed and it is usual to stick with the status quo ante bellum during the RfC. Because it is longstanding, it is unlikely that removal will be an emergency, and so this will be fine. It's a matter of being reasonable in the circumstances.

What I would oppose, however, is any attempt to systematise this by saying that x days/weeks/months is sufficient to establish a "consensus" for the material. What is important is not how long it has been there, but how many people have reviewed it, and at what depth. Sirfurboy🏄 (talk) 19:27, 29 April 2026 (UTC)

What is important is not how long it has been there, but how many people have reviewed it, and at what depth. Is this not a much-needed clarification we should make to WP:ONUS? I'm fine if we go another route than saying X days/weeks/months is sufficient to establish presumed consensus in order to limit ONUS' scope (like basing presumed consensus on the number of page editors instead of on duration), but whatever route we go should be a proposed change addition because clearly the current wording is not cutting it. Alexandraaaacs1989 (talk) 20:07, 29 April 2026 (UTC)

I don't think there's anything wrong with the wording. Wikipedia is not a bureaucracy, and every rule is mediated by contact with the reality of each situation. When it comes down to it, every edit war (where participants are not blocked) is resolved when people start talking. Onus requires people to start talking. Sirfurboy🏄 (talk) 20:44, 29 April 2026 (UTC)

I'll mention here please do not quarantine other editors' comments with no consensus on public discussion forms just because you do not agree with them. You do not unilaterally get to decide what is/isn't relevant to this discussion.

When it comes down to it, every edit war is resolved when people start talking. Onus requires people to start talking. WP:CON also requires people to start talking. If making people talk is the goal, why do we need WP:ONUS when CON already achieves this exact goal? All this does is give deletionists the capacity to drive-by ONUS delete without having to follow regular consensus-building rules that apply to other editors. Alexandraaaacs1989 (talk) 22:14, 29 April 2026 (UTC)

Hatting off topic parts of a discussion is normal. See WP:SUPERHAT. You reverted my hatting so I won't edit war it, but that section is off topic. You want to talk about me (obliquely) and WP:BIDIRECTIONALity. You have threads elsewhere talking about these things, and there is a thread at the template page itself that you declined to comment on. You are WP:FORUMSHOPPING. I won't discuss bi-directionality with you here, and the section needs hatting. It is off topic. Sirfurboy🏄 (talk) 22:39, 29 April 2026 (UTC)

None of our policies actually require discussion from any particular person. You can even refuse to participate in an ArbCom case. What we require is that if the situation reaches a certain point, then someone (or a specific person) has to start (or participate in) a discussion or live with the consequences (e.g., inability to restore information in the article, getting blocked because nobody could understand why you had done the things you did before, etc.). WhatamIdoing (talk) 23:25, 29 April 2026 (UTC)

I haven't added my voice to this particular ONUS discussion but I think it's pretty straight forward. ONUS should never be a reason to remove content. Rather ONUS says unless we had previous consensus, the challenged material needs to stay out until consensus to include is established. Using the Trump hair dye example, if that information was just added and immediately challenged then both NOCON and ONUS would say it stays out until a consensus to keep has been established. What about the case where the material was added a while back and never noticed (perhaps the article has little traffic or it was lost when a lot of other edits were made)? In that case the person removing the material would argue, in effect, the material was UNDUE. Hopefully with something more than just "UNDUE". Once it was removed, it shouldn't be restored per ONUS until a consensus for inclusion is reached. The original removal should not justify the change with "ONUS" as that's like justifying the original removal with "NOCON". It doesn't say why the material should go. Also, ONUS applies to things that can be WP:V. It doesn't apply to changes that aren't specifically content related. Springee (talk) 23:53, 29 April 2026 (UTC)

Do you disagree that ONUS should be updated to better reflect the nuance you just described? Because I think what you just said should be codified. Alexandraaaacs1989 (talk) 00:39, 30 April 2026 (UTC)

I'm fine with a clarification to that end. I see that as not changing the meaning/intent. Springee (talk) 01:01, 30 April 2026 (UTC)

Oops, I misunderstood what you were saying. I think we actually disagree.

I am curious how situations where information in a high-visibility article for two years being suddenly ONUS-deleted for any reason at all should be handled. Isn't there a presumed consensus in some situations that means the content should be included rather than removed until the dispute is concluded? And shouldn't the person challenging the status quo be the one forced to start the talk discussion arguing how the material with presumed consensus does not deserve a presumed consensus? We need a rule to account for drive-by ONUS-deleting in some capacity. The rule is too abusable in its current form. Alexandraaaacs1989 (talk) 05:52, 30 April 2026 (UTC)

What even is your oft repeated "drive-by ONUS deleting"? How does this differ from normal editing? Suppose (and this is based on an actual case I found) that an editor was reading an article about a well known book, and they discovered that two names had been added to the cast list that were not in the book. Suppose that the editor, familiar with this book, having just read it, realised - checking the history - that the two cast names of supposed children in the book were added by an IP address suggesting an insertion from a US middle school. The editor removes the names, but even though they had been inserted months before, the IP puts the names back shortly thereafter. The page has many other edits, but other editors had not noticed this small change. Was the removal a drive-by? If that editor removes the names again, and cites ONUS (leaving a friendly message for the schoolkids as they do so), is that inappropriate?

Last week I corrected something in an article that had been inappropriately edited 3 years ago, and no one noticed the sentence made no sense. Editors do not read the whole of articles they watch every time they edit them. Things get missed. Fixing errors when they are found is not drive-by anything, it is just editing. We don't need a rule to make it difficult for people to edit erroneous content. Sirfurboy🏄 (talk) 07:38, 30 April 2026 (UTC)

But what about when the drive by removal of information isn't erroneous content, but is actually well-sourced and WP:DUE, etc? Perhaps someone wants to remove content that has had long-standing consensus for inclusion, but due to their own POV or political tastes, decides to take it out of the article anyway? Shouldn't the responsibility fall on the deletionist editor to make a case for removal? ~2026-24524-49 (talk) 17:58, 30 April 2026 (UTC)

I cannot think of any such case I have seen where the removal was not reverted with an edsum or comment to the effect of "see the consensus at..." Once consensus is established, ONUS cannot be used to keep information out. If, however, you are asking about something that has no consensus but is correct, then the consensus should quickly follow. But the caution here is that sometimes what seems obviously right may not be so. That the process gets the information checked and restored with consensus if right is a positive thing. Having it missing from the article for the brief period it is being discussed is unlikely to be an emergency. Sirfurboy🏄 (talk) 18:08, 30 April 2026 (UTC)

But there isn't always a formal RfC to point to that shows consensus was established. Sometimes consensus occurs informally, through the normal course of edits which have been made and refined by numerous editors over time, etc. ~2026-24524-49 (talk) 18:32, 30 April 2026 (UTC)

If a number of editors refined a block of text then that mutual edit process may indicate an explicit consensus. That can always be checked by pinging those editors to a discussion related to the disputed material. Pinging editors who previously worked on the content isn't canvassing, it's just typical notification. Springee (talk) 19:34, 30 April 2026 (UTC)

If a mutually constructed block of text is itself indicative of consensus and therefore is a situation where ONUS cannot be invoked (I agree), then we should edit ONUS to clarify that mutual edits are a form of consensus and therefore consensus for deletion must be created before deleting content that was mutually edited. We should also explore edge cases where it's ambiguous whether a block of text was mutually updated if we make this change. Alexandraaaacs1989 (talk) 22:40, 30 April 2026 (UTC)

I think many editors (including myself) would oppose any change that would require one to dig through article history or get pre-clearance in some form to delete something from an article. BRD is plenty good enough for that situation. MrOllie (talk) 22:45, 30 April 2026 (UTC)

Similarly, nobody wants to have to dig through article history to defend it and prove that some text has informal consensus for inclusion due to long-standing mutual edits by numerous editors. There has to be a better way than all this. ~2026-24524-49 (talk) 01:08, 1 May 2026 (UTC)

That's why I originally proposed the 1-week-since-edit threshold for whether ONUS can be applied. I think it keeps things much simpler that way. At minimum, I think the 1-week threshold should apply to pages designated as covering contentious topics and it may not be necessary elsewhere. Alexandraaaacs1989 (talk) 01:13, 1 May 2026 (UTC)

I'd rather people just discuss whatever the underlying content issue is than debate whether or not edits are mutual or whether or not an article is contentious enough that some special threshold applies. MrOllie (talk) 01:32, 1 May 2026 (UTC)

There has to be a way to make this happen without leaving the door open anyone to take down any content for any reason without consensus and citing ONUS when others try to add it back. Alexandraaaacs1989 (talk) 03:25, 1 May 2026 (UTC)

The original editor should have valid concerns for removing the content. If those concerns are valid editors shouldn't add it back without resolving the issue, if the concerns are not valid and an editors continues in that way then it's a behavioural issue. -- LCU ActivelyDisinterested «@» °∆t° 16:33, 2 May 2026 (UTC)

If enough different editors are putting the content back, then the "one against many" situation is a sign of consensus by itself. WhatamIdoing (talk) 05:16, 3 May 2026 (UTC)

In re But there isn't always a formal RfC to point to: An WP:RFC is not the only way to establish consensus. WhatamIdoing (talk) 05:13, 3 May 2026 (UTC)

Onus should never be the justification for removing the content, it's only the reason why the content shouldn't be restored. If the content was in the article for a day or a decade, ONUS isn't the justification for removal. Assuming the material passes WP:V, then removal should boil down to a question of weight. A claim from a low quality source has low weight. A claim that is not relevant to the article/article section has low weight (or BALASP). When the material is removed the justification should be something like "This is UNDUE because...". At that point, per ONUS the material should not be restored unless a prior consensus for inclusion can be shown. If the removal was bad it will likely be overturned but the consensus to keep will have to be shown. If we are dealing with a new addition then this is the same as NOCON. Like ONUS, NOCON shouldn't be the initial reason to remove something. It's a reason why something shouldn't be/shouldn't have been restored. The difference is ONUS is limited in scope (claims that pass V but haven't shown explicit consensus) but unlimited in time (recent or not recent). NOCON is unlimited in scope (content, style, layout, picture etc) but it's limited in time. So a paragraph that was moved last week vs last year is different.

Part of the value of ONUS is dealing with content that was missed when originally added (low traffic, lost during a bigger set of edits). In effect, ONUS cancels implicit consensus and forces an explicit consensus or a removal. That was longer than I intended but hopefully clarifies things. I feel like it needs a for chart. Springee (talk) 19:28, 30 April 2026 (UTC)

To say ONUS is limited to scope when it covers everything that isn't already covered by something else (failing WP:V) is a bit of a misrepresentation. It applies to everything except.. content that it doesn't matter whether or not it applies to. Also, of course ONUS isn't a reason to remove content itself, but that doesn't mean much considering nearly every claim that someone would want to remove from the article is also something that a plausible sounding argument could be given against, especially by citing some form of WP:NPOV violation. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 20:03, 30 April 2026 (UTC)

I see ONUS only covering content, not style, formatting etc as they aren't someone covered by WP:V. Perhaps a different way to say it is ONUS tells us how to handle a removal of a fact and who is responsible for showing a consensus to include once something has been challenged. Springee (talk) 21:40, 30 April 2026 (UTC)

Apologies if this is not the proper place to post this. Years ago, it seemed common for articles to include links to Wikimedia sister projects such as Wiktionary or Wikisource (whichever was appropriate for a given article); nowadays I rarely, if ever, come across them. Did the community decide, at some point, to ban that practice, for reasons? If so, is it worth revisiting? I found such links very useful, personally. Or, am I just imagining this change of trend? Lastly, if such links are still allowed, is there a template for adding them? Thanks, ~2026-23890-25 (talk) 08:08, 27 April 2026 (UTC)

The interwikilinks are replace by Wikidata. The Wikidata item link on the side of the page shows not only data on the topic, but links to other wiki language and project pages. Graeme Bartlett (talk) 08:15, 27 April 2026 (UTC)

For most people, it's not in the side bar. It's in the "Tools" menu, which is a drop-down menu near the "View history" tab. WhatamIdoing (talk) 02:08, 29 April 2026 (UTC)

They still exist. For example, Treaty of Tientsin has a link to the treaty's text on Wikisource in the bottom right corner, while many train articles like British Rail Class 231 have links to Commons in similar boxes. Toadspike [Talk] 08:40, 27 April 2026 (UTC)

There's a practice in Deaths in 2026/Recent deaths to specifically delete interwiki after 30 days if en wiki doesn't have an article, which I think is atrocious. Jahaza (talk) 02:19, 29 April 2026 (UTC)

 Interested editors are invited to join the discussion at Wikipedia talk:WikiProject Classical music § RfC: amending project guidelines on infoboxes. MichaelMaggs (talk) 21:16, 27 April 2026 (UTC)

If I may suggest, I would like to raise a possible clarification on NPOV and due weight for allegations and controversies.

NPOV asks us: "carefully and critically analyze a variety of of reliable sources" and present views: "fairly and proportionately" so content reflects the relative prominence of views in those sources. Policies like NPOV, NOTSOAP, PROMO, NOTSCANDAL all point in same direction. Meaning, it's not just "is this verifiable?" but also "how much weight does this carry"? in practice, something can pass the "it's in a reliable source" test, but not the weight test. For example, a claim in a single source can make its way into a wikipedia page even when there hasn't been much follow-up or wider coverage. This is especially relevant to serious allegations, where limited or one-off reporting without broader or continued coverage can give more weight than the sources support. It can look bigger on the page than in the wider world of sources.

We can see that we have clear guidance for promotional or advocacy content like WP:PROMO, which helps keep things from getting too heavy on the positive side. The flip side, how to handle lightly covered negative or controversial content relies more on general due weight rules. So, the question is: when something is verifiable but minimally covered with possibly no follow-up or wider lasting coverage or updates, how should we decide if it has enough prominence to include without giving undue weight?

One way might be to clarify this with something like this: "Allegations or controversial content is generally appropriate when supported by significant coverage in independent, reliable sources. Content based solely on isolated or limited reporting or single sourced one-off reporting, even if verifiable, may be excluded if its inclusion would give undue weight. The presence of a denial does not by itself establish due weight". This would help illustrate how due weight applies in a common edge case. If is notable enough for inclusion, it would have atleast a follow-up or an investigation or multiple sources with some in-depth reporting and this clarification will not impact inclusion of such content or any discussion around such content.

Curious what others think. Would something like this be useful, perhaps as an example within WP:UNDUE or WP:PROPORTION? NicoR8 (talk) 17:55, 28 April 2026 (UTC)

We have stronger protections for BLPs, which covers the vast majority of problematic, poorly-verified allegations. I think you're correct that for other topics, like companies, the rules are just due weight. I can imagine situations where this would be less than ideal. Generally editors use their common sense to reach reasonable outcomes, though, so I'd have to see evidence that this is a real and persistent problem before considering modifying policy. Toadspike [Talk] 19:54, 28 April 2026 (UTC)

@Toadspike Yea that makes sense and I suppose commonsense plays an important role and I saw BLP covers many of the more sensitive cases and that is what placed me in dilemma about this edge case in non-BLP situations like as you said companies or deceased people. For example, someone's claim might be attributed, but without much follow-up or wider coverage (Talk:Paul Crouch#Question about BLP Policy) it's veriifiable, yes, but how much weight it should carry isn't always obvious.

That's when things can get a little wobbly. Editors can look at similar content and land in different places (Wikipedia:Neutral_point_of_view/Noticeboard#Ron_Wyatt per one editor " it seems to be in reasonable shape" and Special:Diff/1351575584 - per another editor "a bunch of unacceptable sources. Is there any secondary coverage of this" ), same ingredients, different pot. Another example Wikipedia:Neutral_point_of_view/Noticeboard#Millsaps College article – section weight and neutrality is waiting for help since April 10 ( might not be quite related ), and would have helped if some NPOV/UNDUE clarification was available.

A small clarification or example(s) based on what policy already says, not a new rule, might help steady things in these edge cases. Just making the guidance a bit clearer so we don't end up going round in circles. Of course, like any policy, it would still rely on commonsense, but with a bit more consistency and some example-based thinking to guide it, so editors are a bit more aligned in how to apply the same due weight rules consistently. For the kind of cases you thinking, an example like that could help make the guidance clearer in practice. NicoR8 (talk) 11:58, 29 April 2026 (UTC)

I think that two types of examples could be useful to editors:

• Avoid "criticism" and "reaction" sections. The goal is for someone reading about the subject to get both pro- and anti- sentiments in the same place. Otherwise, some readers will see only the main section and not see the criticism, and other readers will skip to the criticism section and never see the supportive information. For example:
  • When writing about a creative work, business, or product, try to put all related things into the same section (e.g., ==Products== or ==Corporate history==): "In 2026, Big Business, Inc. produced a cheaper version of their popular widget, which was criticized for having lower quality." Don't put "In 2026, Big Business, Inc. produced a cheaper version of their popular widget" in the ==Product develompent== section and then have a separate ==Criticism== section that says "In 2026, a cheaper version of their main product was criticized for having lower quality".
  • When writing about an event or politics, try to put al related things in one section (e.g., ==Recommendations== or ==Campaign themes==): "This report recommended changing the initial requirements, which was criticized for excluding people and not following scientific standards." Don't put "This report recommended changing the initial requirements" in one section and then put "The report was criticized for excluding people and not following scientific standards" in a separate ==Reactions== statement. (If the reaction is perfunctory or generic (e.g., a politician offering Wikipedia:Thoughts and prayers after a disaster), then it should normally be omitted.)
• Only include criticism when it is significant or unexpected. For example:
  • Omit Captain Obvious, like "Demicans disagreed with Republocrats".
  • Try to write for your opponent, and especially to search for sources that support other viewpoints.
  • Think carefully about the difference between a viewpoint being popular, especially within a filter bubble, and a viewpoint being presented in high-quality sources.
  • etc.

The second point could probably be expanded into half a dozen useful essays. WhatamIdoing (talk) 20:39, 29 April 2026 (UTC)

Please consider joining the feedback request service. An editor has requested comments from other editors for this discussion. This page has been added to the following list: Wikipedia policies and guidelines When discussion has ended, remove this tag and it will be removed from the list. If this page is on additional lists, they will be noted below.

Should Wikipedia:Presumptive removal of AI-generated content be promoted to a guideline? (see here for prior discussion) Kowal2701 (talk, contribs) 20:29, 28 April 2026 (UTC)

Support per prior discussion, checks to prevent abuse of the process are satisfactory and this would help relieve some of the burden on AI content cleanup. Chaotic Enby (talk · contribs) 20:34, 28 April 2026 (UTC)

Support as one of the proposers, this is necessary because of the asymmetry between the little time and effort required to produce AI-generated content, and the massive amounts of time and effort currently needed to clean-up said content. I hope that any small improvements !voters may have can be uncontroversially carried out after the RfC if adopted. Kowal2701 (talk, contribs) 20:36, 28 April 2026 (UTC)

Support per Chaotic Enby. --Gurkubondinn (talk) 20:40, 28 April 2026 (UTC)

Support - I've been following the discussion (though not really commenting), and I think it makes a lot of sense. The proposed guideline seems well written and covers most edge cases. I agree with Kowal2701 that something like this is needed because of the asymmetry. InfernoHues (talk) 20:41, 28 April 2026 (UTC)

On the understanding that this guideline could be developed and refined in the normal way for guidelines after implementation, I would support.—S Marshall T/C 20:50, 28 April 2026 (UTC)

Support. This will be a useful tool against large-scale LLM abuse. I'm glad much feedback, including my own, was incorporated during the earlier discussion. Toadspike [Talk] 20:52, 28 April 2026 (UTC)

Support - A necessary improvement to reduce AI-related cleanup. Rambling Rambler (talk) 20:55, 28 April 2026 (UTC)

Support Something like this is needed, and the safeguards seem fairly well thought out. Meters (talk) 21:00, 28 April 2026 (UTC)

Support - go for it. GoodDay (talk) 21:02, 28 April 2026 (UTC)

Support as semi-kind-of-informal-co-proposer. Fermiboson (talk) 21:44, 28 April 2026 (UTC)

Support I floated something like this at WT:AIC back in November, so am happy to see this (much better) version proposed now NicheSports (talk) 21:50, 28 April 2026 (UTC)

Support. -- sapphaline (talk) 21:51, 28 April 2026 (UTC)

Support and not a minute too soon. I'm really glad to see these proposals are still being worked on and brought forward - well done all involved. Andrew Gray (talk) 22:22, 28 April 2026 (UTC)

Support lgtm Feeglgeef (talk) 22:46, 28 April 2026 (UTC)

Support: I'm not the biggest fan of the way the first bullet point of the Requirements section is phrased, but it gets the job done, and the sooner we get a functional guideline like this into place the better. Figuring out he most elegant way to phrase things can be done afterwards, but right now this is seriously needed to address the fact we live in an age where the means to mass produce slop have been given to everyone with an internet connection. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 23:58, 28 April 2026 (UTC)

Support, although I think WP:LLMPRVOBJ should we tweaked. Once content has been deleted it shouldn't be restored without achieving consensus first. Paprikaiser (talk) 00:45, 29 April 2026 (UTC)

I was first skeptical because the article deletion part of the guideline has no teeth and is more-or-less the same as WP:PROD, just with being allowed to start an edit war about whether the tag should be in the article. But "edits to existing pages may be reverted without review", similar to the presumptive deletion of copyright violations, would be a good thing to have as part of a guideline or policy. I don't think it needs its own guideline page, though. ~ ToBeFree (talk) 00:57, 29 April 2026 (UTC)

Support It seems good enough; if problems arise when it is applied in practice, we can address those then. Stepwise Continuous Dysfunction (talk) 02:44, 29 April 2026 (UTC)

Support. At least now there will be a concrete guideline to point to when people inevitably get yelled at and scolded for doing this. Gnomingstuff (talk) 03:04, 29 April 2026 (UTC)

Support as proposed 〜 Festucalex • talk 03:31, 29 April 2026 (UTC)

Support. This has deserved to be an official guideline for far too long. 𝔰𝔥𝔞𝔡𝔢𝔰𝔱𝔞𝔯 ^{(𝔱𝔞𝔩𝔨)} -⃝⃤ (they/he) 05:03, 29 April 2026 (UTC)

Support based, as Kowal puts it, on the "asymmetry" between the time need for LLM generation and human cleanup. Assuming LLM tech advances in its plausibility and agentic AI gains broader adoption, that asymmetry will only increase. Rjjⁱⁱⁱ (talk) 08:37, 29 April 2026 (UTC)

Support per Kowal2701, who is paraphrasing Brandolini's law. ~~ AirshipJungleman29 (talk) 15:42, 29 April 2026 (UTC)

Oppose: Nobody wants this work to be done. People will yell at you for doing it. There is no consensus for this guideline as long as people will yell at you for even just tagging an article and not deleting any content -- and in fact will accuse you of nominating the content for deletion, even when you didn't -- and so we shouldn't set people up for disappointment by telling them it's OK to do something that people don't think is OK and will give them endless grief for. I'm so tired. Gnomingstuff (talk) 16:46, 29 April 2026 (UTC)

I take your concerns seriously, believe me, but I actually find your struck-out comment above more persuasive. People sniping at each other and escalating disagreements into drama is, I think, sadly inevitable. But having a solid guideline to point to may, at least, tilt the balance in favor of dispute non-escalation. That's kind of the point of many guidelines, secretly or not-so-secretly: to stop arguments early, or to stop the Nth iteration of the same argument. Stepwise Continuous Dysfunction (talk) 17:37, 29 April 2026 (UTC)

I no longer have any faith that a solid guideline will make any difference when people can just pretend it doesn't exist. Gnomingstuff (talk) 17:40, 29 April 2026 (UTC)

Fair point. I won't try to talk you out of your oppose !vote, just to maybe clarify my thinking in case that is helpful to anybody. I agree that some amount of people doing what you describe is unavoidable; it has happened with every other guideline too. My hope (and gosh how unusual it is to use that word these days) is that when the inevitable arguments start, a concrete guideline will help them end faster. Stepwise Continuous Dysfunction (talk) 18:14, 29 April 2026 (UTC)

This is what sanctions are for Kowal2701 (talk, contribs) 18:22, 29 April 2026 (UTC)

@Gnomingstuff, I'm a bit concerned that you're letting your personal experience jade you here. I do recall that you've been quite publicly criticised (and dragged to ANI, if I recall) in the past for your 'drive-by tagging' of articles as AI generated, but this isn't the first time now that I've seen you bring up that incident like an old war scar in discussions about AI policy to support a kind of doomerist position.

It's unfortunate you had that experience, and it's unfortunate that there's always going to be people who lash out instinctively when their content is tagged as 'bad,' but I think the recent broadening of WP:NEWLLM into WP:NOLLM and the mass support for proposals like this is a great sign that the attitude towards AI on Wikipedia is increasingly becoming more and more negative.

There's always going to be people who don't read guidelines and then get mad when people enforce them; what's great about having those guidelines is that you can immediately invalidate their arguments by saying "My removal of your content was perfectly justified by WP:LLMPROD." Isn't that a luxury you would've liked to have had when you were being raked over the coals for no good reason? Why does your experience support the idea that we shouldn't give more explicit support to people who end up in your very position? Athanelar (talk) 00:05, 30 April 2026 (UTC)

This is about a more recent incident, where I had WP:G15 (the speedy deletion policy) quoted at me -- by someone who was not even the original editor -- to "explain" to me why AI-generated text was "not a major problem," and then dragged me to the NPP review board to further scold me for it. Did I tag the article for G15? No. Did that make any difference? Also no. Did bringing up our actual guidelines "immediately invalidate their arguments"? Apparently not, and so the article that is almost certainly AI-generated (considering the original editor has indeed admitted to using AI on other edits) remains untagged. When there is this much opposition, this regularly, producing this much of a chilling effect for just tagging content, how are we even talking about deleting content? But yes, the case you bring up would be another example. Gnomingstuff (talk) 00:12, 30 April 2026 (UTC)

plus in the discussion below (which I literally don't know if I'm allowed to contribute to) you have stuff like if contested I would expected any attempts to re-add include some AISIGN in the article itself. -- if there is not even buy-in to presumptively tag articles then how the hell are we even talking about deletion? Who is this person who apparently arbitrarily rolls the dice on whether they will use AI for any particular article, with motivations that are extremely unlikely for any actual person but whose hypothetical existence we have to hem and haw about forever? Gnomingstuff (talk) 01:56, 30 April 2026 (UTC)

I believe that both approaches are consistent, as both presumptive tagging and PROD offer a way for editors disagreeing with it to contest it and take responsibility for the content. Circumstantial evidence is stronger than no evidence, and removing a tag based on the former on the basis on "I don't like this tag" wouldn't be a solid argument, but doing so on the basis of, say, a change in editing patterns can be stronger (which is already something we consider in this draft guideline!) Chaotic Enby (talk · contribs) 02:00, 30 April 2026 (UTC)

You'll notice, though, that there isn't really any hemming and hawing in the survey right now. You're literally the only oppose voter right now, and it isn't even an oppose reasoning so much as it is just a bitter expression of spite.

The participants of the survey are wholly in favour of the presumptive tagging, reversion and deletion of content added by editors blocked for AI usage or otherwise identified by consensus as being problematic AI users. That's a huge step forward for our position on AI. Athanelar (talk) 02:01, 30 April 2026 (UTC)

What I am arguing is that, for whatever reason, none of these surveys are capturing the actual community opinion in practice. There are always people who come out of the woodwork to suddenly oppose it. Gnomingstuff (talk) 02:06, 30 April 2026 (UTC)

I get that; but if they don't come to the RfC, and this guideline is enshrined into text, then it sucks to be those people because if they later come to complain about a decision made on the basis of this guideline, the complaint is already ipso facto moot.

Consequently, your argument rather seems to sound like "what's the point in making rules if people will complain about them and try to avoid them", to which the answer would seem to be; well, because we can then impose sanctions on those people for breaking the rules. Athanelar (talk) 02:18, 30 April 2026 (UTC)

I mean, look at the topic below "Clarifications on the applicability of NEWLLM". At least two opinions expressed there are incompatible with this, as they oppose even presumptive tagging of AI-generated content. So they might as well be oppose !votes here. Gnomingstuff (talk) 19:36, 1 May 2026 (UTC)

Support. I have nothing much to add argumentatively from what others have said; I was planning to remain uninvolved in this so I could close it promptly when the opportunity arises, but anybody who knows anything about my editing would have more than enough grounds to accuse me of being involved anyway.

I think this is the perfect example of a descriptivist proposal. This strengthens the already-existing processes in use against AI editors. Athanelar (talk) 00:28, 30 April 2026 (UTC)

Support. Having a nuke option would be fantastic, as long as its done extremely sparingly, and only in the most obvious cases. EatingCarBatteries _{(contribs | talk)} 01:23, 1 May 2026 (UTC)

• Support reluctantly – Another rule on AI-generated content is still needed. The Requirements section as-is is intriguing, thought it might change should it become either inadequate or excessive. Also helps combat AI-generated sockpuppetry. --George Ho (talk) 05:05, 1 May 2026 (UTC)
• Support per others who have summed up my thoughts adequately. Regarding Gnomingstuff's concerns, if other editors are screaming at you for tagging an article as AI-written, that's 1000% on them, not you. XtraJovial (talk • contribs) 22:53, 1 May 2026 (UTC)
• Support. We need it as a guideline to provide a point of reference. AI content is far too serious a matter to be settled on a case-by-case basis, especially as those so often become disputes. We must have a consistent approach, which will be enabled by a guideline. I would think the issue will eventually be discussed in policy terms, but for the present a guideline will be a big step in the right direction. Jack (talk) 10:40, 3 May 2026 (UTC)
• Support per my experience with LLM users NotBartEhrman (talk) 11:13, 3 May 2026 (UTC)
• Support aiming to avoid polluting LLMs while ensuring the factuality of wikipedia honestly nhals8 (rats in the house of the dead) 11:40, 3 May 2026 (UTC)

To me, WP:GAME is a bit too subjective, and slightly discourages segmented edits. Segmented edits are better in some cases, as they separate possibly controversial and minor edits. So here is my idea:

1. We create a system where you can make as many edits as you like to a page, but only the first in a given amount of time count. (e.g. you can do five consecutive edits in 20 minutes to Alhambra, but only the first counts.) My first thought would be that main space pages have a "cool down" of 1 hour (this is per page, you could edit one page, then edit a different one, and that would count as two) and non main space pages have a cooldown of 1 day (again, per page.)

2. We encourage segmented edits to separate minor and controversial edits.

3. We change the requirements of edit counts to autoconfirmed, extended confirmed, etc. (Maybe 7 for autoconfirmed, 350 for extended confirmed, and decrease anything else needing edit count by 20-30%) Wikipedian12512(alt) (talk) 16:47, 29 April 2026 (UTC)

That is an enormous amount of complex, hard to implement, controversial changes for a slight gain. In other words. not a chance. --Elmidae (talk · contribs) 17:47, 29 April 2026 (UTC)

The segmented edits is one part of it, but it also improves other areas. If we had a button that disabled inappropriate page blanking, then we would use it, even if it was a "hard to implement, controversial change for a slight gain." Wikipedian12512(alt) (talk) 18:12, 29 April 2026 (UTC)

I'm assuming you mean WP:PGAME? In that case I just don't see the issue. It clearly says that making trivial edits is only an issue if done to raise your user level. Adding a substantial edit in segments where each segment added is substantial on its own is not an issue. Even gnoming isn't an issue. As long as your edits are actually meaningful and not done solely to get user rights it's fine, and I think the guideline reflects that.

Also, while we're at it, what is up with WP:GAME#Various levels of intent? I just looked at it again now while I was on the guideline page and I've got no idea what half of it is trying to say. That has got to be rewritten. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 18:47, 29 April 2026 (UTC)

Readers of this thread might also want to see Wikipedia:Village pump (idea lab) § Something about WP:GAME. Graham87 (talk) 07:14, 30 April 2026 (UTC)

If it's appropriate to make a complex series of changes as a series of small edits, then do so. No one will accuse you of gaming the system. -- LWG ^{talk (VOPOV)} 14:45, 30 April 2026 (UTC)

Exactly. To give another example of how small edits≠permission gaming: I got extended confirmed while doing high-speed manual mass removal of the |protection= parameter from {{Contentious topics/talk notice}} to clear the massive (>500 at the time) backlog of pages in Category:Wikipedia pages about a contentious topic mislabelled as protected. That's not gaming the system because such mass edits are necessary to fix the problem. No one minded, and I even got my first barnstar for this. SuperPianoMan9167 (talk) 17:23, 30 April 2026 (UTC)

I'm not going to put an RfC tag on this because that would be an utter waste of everyone's time, but I want to pose two questions to the community.

1. Is circumstantial evidence sufficient to tag an article as LLM-generated? (For example, if an editor has demonstrably used LLMs on talk pages and past articles and has failed to respond to LLM warnings, and writes a new article, can one tag said article as LLM-generated based on the prior evidence without citing AISIGNS within the article itself?)
2. Is LLM-generated content without any policy violations otherwise (a) allowed in mainspace and (b) not to be tagged?

I believe the community currently does not have a consensus on these two questions, based on a number of threads which I will not link because I do not want to reexcite their activity. My personal analysis, based on preliminary off-wiki discussions with a number of editors I am fairly confident will remain out of this discussion, is as follows:

1. Yes. The tag is not a verdict in a criminal court. If one is unable to find AISIGNS within the article, that decreases the probability the article is AI generated; the circumstantial evidence increases the probability the article is AI generated. If the latter is sufficiently strong to overcome the former, then the tag should be placed until such time as the article is sufficiently edited by other editors (or by the same editor in a clearly human manner, etc) that the probability that the text is AI generated is sufficiently low.
2. (a) No, (b) no. WP:NEWLLM states that LLM-generated content is prohibited in Wikipedia, period. It does not make exceptions as to whether the content is reviewed or as to the quality of the content. The tag says "may contain hallucinations, etc, etc" and does not say "does" (though some thought may be given to changing the wording, I don't see how it could be much improved). We could reach a new consensus as to this question, but that would require changing the text of NEWLLM.

I welcome comments below. Fermiboson (talk) 19:06, 29 April 2026 (UTC)

imo, re 2) the spirit of it in its current state would be "yes" since it's predicated on content violating PAGs. "No" would mean an ideological ban, which I was thinking the community would decide on further down the line (it was quite controversial in discussions leading up to the RfC, and the main reason we didn't have a LLM policy for 3ish years). But everyone I talk to seems to interpret NOLLM as a total ban regardless of PAGs violations Kowal2701 (talk, contribs) 19:12, 29 April 2026 (UTC)

My understanding:

1. The only appropriate reason to add the tag is that you believe there is content in the article that is LLM-generated. When tagging you should indicate which edits the concern applies to (ideally with diffs) and at least some indication of why you suspect those edits were LLM-generated. The only appropriate reason to remove the tag is if the LLM text has been removed, or if you believe the edits in question are not in fact LLM-generated. Merely disagreeing that LLM text is a problem (ie disagreeing with WP:NOLLM) is not grounds for removal, just like disagreeing with WP:OR is not grounds for removing that tag.
2. No, LLM-generated content is not allowed in mainspace. LLM-generated content is inherently a policy violation of WP:NOLLM. That guideline's existence is predicated on the frequency with which LLM's violate PAGs, but each individual application of that guideline is not. At the present time this community is simply not equipped to satisfactorily assess the quality of LLM content to a standard that makes it acceptable to include it.

This is my understanding of the overwhelming consensus at the WP:NOLLM RFC. -- LWG ^{talk (VOPOV)} 19:37, 29 April 2026 (UTC)

My understanding:

1. No. Placing the tag on an article is an assertion that you have read the article and as a result of that reading have a good faith belief that it contains LLM-generated text and that text is sufficiently voluminous and/or otherwise problematic that the article needs actively editing. You must be able to identify, when asked, specific issues with the article text that you believe in good faith to be the result of an LLM.
2. Yes. The placement of a tag is an assertion that the article has problems that need editing to resolve. If the text on the page would be unproblematic if written by a human then there are no problems with the article. The community endorsed no-LLM on the basis that such text is more often problematic than not, it did not endorse a moral prohibition (even though some editors would like that to be true). Thryduulf (talk) 19:48, 29 April 2026 (UTC)

My own understanding:

1. Possibly, but similar to any other tag (and including WP:LLMPROD here), it can be contested, at which point you should be able to cite issues specific to that article.
2. No per WP:NOLLM. While theoretically that content may not be problematic, requiring us to accept it puts us back in the same position as before of having to check every single piece of LLM-generated content, bringing up the same issues of asymmetry of effort. Chaotic Enby (talk · contribs) 00:14, 30 April 2026 (UTC)

To resolve the apparent contradiction: the "it can be contested" implies that the editor contesting the tag is taking some level of responsibility for the content, and has checked it on their own, which means we aren't putting the burden on the patroller. While circumstantial evidence is weaker than direct evidence from the article, it is still stronger than no evidence at all, meaning the contesting editor should have some specific reason to believe that the edit hasn't been AI-generated (lack of WP:AISIGNS, spotchecks, etc.). Chaotic Enby (talk · contribs) 01:53, 30 April 2026 (UTC)

A good way to clarify this might be to say in the guideline that all LLM text in articles is presumed to violate our core content policies unless otherwise demonstrated by a contesting editor. I₂Overcome ^talk 05:11, 1 May 2026 (UTC)

That is great! Chaotic Enby (talk · contribs) 05:14, 1 May 2026 (UTC)

Do you think it would be worthwhile to start an RfC to potentially add something like this? I₂Overcome ^talk 18:40, 1 May 2026 (UTC)

Not really. It's already quite implicit, and I'm worried that adding this in the text would open the wikilawyering door to any author of LLM text saying "it doesn't violate our policies, I verified it!" (despite this verification virtually always being superficial at best), which was already a common occurrence before, and would put the burden back on patrollers. The much stricter criteria of WP:LLMPROD (already up to RfC right now) do this better. Chaotic Enby (talk · contribs) 18:52, 1 May 2026 (UTC)

But there are at least two experienced editors here who disagree that NOLLM allows patrollers to tag or remove LLM text based on AISIGNS alone, without having to check for policy violations. LLMPROD is a great step in that direction, but it won’t easily apply when, for example, a patroller encounters a user who wrote just one AI article or made a couple of AI edits. In those situations, it would be helpful to be able to presume that the content is problematic without having to manually verify everything, get consensus at AINB, or wait for the editor to be blocked, become inactive, or admit to LLM use. I₂Overcome ^talk 21:10, 1 May 2026 (UTC)

Well, if WP:NOLLM can't be enforced by itself and just means that LLM may violate other policies, then it would be a bit of an empty guideline? But that's not what the guideline says. It says For this reason, the use of LLMs to generate or rewrite article content is prohibited: the prohibition originates from the fact that they often violate existing P&Gs, but nowhere is it said that their use is only prohibited when it violates P&Gs. WP:LLMPROD specifies an enforcement mechanism, but the prohibition is already here. Chaotic Enby (talk · contribs) 21:21, 1 May 2026 (UTC)

I agree. I just wish it was a bit more explicit that people doing AI cleanup work only have to suspect that something was AI-generated, not necessarily have evidence that it violates other policies, since other editors apparently disagree. I₂Overcome ^talk 21:32, 1 May 2026 (UTC)

There are some who think that even if an article is probably LLM-generated, it should be allowed to stand unless someone can specifically point to hallucinations, promotional language, etc. in the article in question. I₂Overcome ^talk 21:37, 1 May 2026 (UTC)

(I hope it is OK for me to reply here, apologies if not) The problem here isn't the users who wrote just one AI article or made a couple of AI edits, those cases are easy. The problem is the users who made 500+ edits, some of which are obviously AI, but not all of which are as obvious. (This user who made a enormous amount of AI rewrites to articles in 2025 comes to mind.)

You have to go through all 500+ edits regardless, but without presumptive tagging, you also have to come up with 500 in-depth, bespoke cases for each one, despite it being common sense that if someone made an obvious AI edit on April 20, and another obvious AI edit on May 20, with other obvious AI edits at either margin, then any similar edits in between are probably AI too. And it is, in my experience, difficult to impossible to persuade people who have only looked at one article and not the other 499. Gnomingstuff (talk) 22:09, 1 May 2026 (UTC)

The only thing I can think of is use AINB more often for editors with say 100+ such edits, and just point to the report there. We could also have an essay on presumptive tagging which explains the reasoning behind it (saving you from having to rehash things every time). Another thing, people probably can't tell you're our resident expert on LLM use just from your userpage and misjudge your competence Kowal2701 (talk, contribs) 23:46, 1 May 2026 (UTC)

I think that WP:LLMPRV will help a lot with that. The concern I was expressing was not about presuming something is LLM-generated based on the user’s other edits, but about presuming that something is problematic solely because it appears LLM-generated. In other words, I think we should explicitly state that if an editor suspects LLM use based on several non-policy violations like heavy use of em-dishes, unnecessary bolding, common LLM words, etc., it is okay to tag and/or remove it without wasting time trying to verify the content. I agree with Chaotic Enby that WP:NOLLM implies that it’s okay, but some editors think it’s only acceptable if you can point to a specific policy violation. I₂Overcome ^talk 00:12, 2 May 2026 (UTC)

I think the key thing is that if you tag an article you haven't individually reviewed, you can't then just reflexively reinstate the tag if someone removes it from that article specifically (i.e. they aren't mass untagging) because it implies that they have reviewed that article. If you disagree with them that's fine, but only if you can explain why with reference to the content of that specific article.

If someone is mass untagging that's a different issue. There could be a few reasons for it (some good faith, some not) but assuming it wasn't an obvious error on your or their part it should probably be discussed on your or their user talkpage or at a noticeboard rather than at an individual article level. Thryduulf (talk) 01:12, 2 May 2026 (UTC)

Agreed (though it’s more of an issue if it’s the author quietly detagging) Kowal2701 (talk, contribs) 01:14, 2 May 2026 (UTC)

Usually the only thing it implies is that they don't like a tag on their article. Gnomingstuff (talk) 02:43, 2 May 2026 (UTC)

Firstly that's an massive assumption of bad faith, and secondly you cannot tell that without looking at the individual article. Thryduulf (talk) 08:24, 2 May 2026 (UTC)

No, it's just literally the only thing you can assume from someone removing a tag, that they don't like the tag being there. Anything else is reading things in. But I'm glad that you are still twisting my words into their worst possible interpretation. Gnomingstuff (talk) 12:41, 2 May 2026 (UTC)

There is no twisting of words involved, and assuming that the only reason someone removes a tag is because they don't like it is an assumption of bad faith. It might be that they don't like it but equally it might be that the tag was wrong (or of course both). Unless you look at that article specifically then it is impossible to know which it is. Thryduulf (talk) 13:10, 2 May 2026 (UTC)

What I am saying is that the only thing one can assume by removing a tag is that they don't want it there. Just like the only thing one can assume by adding a tag is that they do want it there. Or like the only thing one can assume about someone removing a hat is that they don't want to wear a hat, and vice versa. There may be other motivations, but only one can be assumed. What part of this is difficult to understand? Gnomingstuff (talk) 14:25, 2 May 2026 (UTC)

What I don't understand is why you continue to feel the need to assume bad faith of someone removing a tag? Thryduulf (talk) 14:28, 2 May 2026 (UTC)

I am not assuming bad faith. I am assuming nothing. You are the one assuming that all tag removals take place in a fantasyland where everyone involved puts entire days' worth of archival research into the job, instead of the more common reality where an article creator silently removes the tag, added by the evil harpies who condemn them to having a template on their article, while changing nothing at all. Gnomingstuff (talk) 14:32, 2 May 2026 (UTC)

This only works, in my opinion, if we establish consensus that removing the tag actually means they’ve reviewed the article, and systematic removal of tags without review (which would be evidenced by for example a history of leaving PAG violations on the page) is considered disruptive. Otherwise we get back to the same issue of asymmetric effort. Fermiboson (talk) 09:44, 2 May 2026 (UTC)

Removal of the tag is an assertion that it is incorrect. There are multiple possible reasons for that, including

• The page has never contained AI-generated material
• The page has been rewritten and any AI-generated material has been removed
• The remover has reviewed the page and believes in good faith for the current revision to be free of AI-generated material (whether previous revisions were or not is not relevant)
  • It is equally possible for this to be a correct or incorrect belief.
• The remover doesn't understand what removing the tag means
• The remover is acting in bad faith

Assuming (as explicitly stated in my comment) that this is not a case of mass removal, then the only way to determine which case applies is to look at the specific article. If you think the removal is incorrect you need to be able to explain why you think that with reference to the specific article. This is not unfair asymmetry, it is a combination of taggers being human, detaggers being human, the need for all parties to assume good faith, and the community consensus that articles containing LLM-generated text can be made compliant through editing to remove that material. Thryduulf (talk) 13:21, 2 May 2026 (UTC)

• The remover doesn't understand what removing the tag means

But what if the remover doesn't understand or believe you after you have explained it, then what?

I'm asking in good faith, because this has happened to me multiple times. --Gurkubondinn 13:37, 2 May 2026 (UTC)

That's when you get other editors involved. What the best course of action is depends on the situation, e.g. why aren't they understanding. It could be that it just needs explaining in a different way, it could be that they don't believe one person but will believe two or three, it could be that there is a CIR and/or language issue, it could be that there isn't actually a misunderstanding but a disagreement (and it could be either one, neither or both of you that is right) etc. The best way forwards for everybody in all these cases is to get multiple opinions on the matter.

However this is beside my point here, which is that your explanation of why you believe the specific article contains LLM-generated text needs to reference that specific article. Saying "The article must have LLM text in it because these other articles do" in response to "But I can't see any LLM text in this article?" is never going to be helpful. Thryduulf (talk) 13:51, 2 May 2026 (UTC)

Last time, I did all of that and they still didn't believe me or anyone else, and the tag was eventually removed. But I do think that "this editor has used an LLM for everything else" is not an unreasonable agreement (but I did not even mention this last time this happened, I just think it is a reasonable argument).

Because LLM (mis-)use is also a behavioural problem, it is not (only) a content problem. It can also be based identified on behavioural clues, so it should follow that it is also reasonable to point to edit behaviour to support a tagging. --Gurkubondinn 14:01, 2 May 2026 (UTC)

What you are describing in your first paragraph is a behaviour issue, that's not something that can be dealt with by tagging. Follow the dispute resolution process as with any other behavioural issue - LLM use does not make it special in any way.

Behaviour clues are fine to use as a reason to presumptively tag an article. However if someone challenges the tagging on a specific article that must be dealt with based on the content of that specific article. If the content of an article is indistinguishable from content written by a human then arguing over whether it is or isn't human is a waste of everybody's time and energy. Thryduulf (talk) 14:27, 2 May 2026 (UTC)

As opposed to the very productive use of one's time that is endlessly haranguing and obstructing people who do AI cleanup. Gnomingstuff (talk) 14:33, 2 May 2026 (UTC)

This is not obstructing people who do AI cleanup, it's simply making sure that AI cleanup does the minimum amount of collateral damage. I get the impression that you are getting burned out. I suggest taking a few days away from the subject of AI/LLMs completely and come back refreshed. It will make you much more productive and significantly lower the risk of you doing or saying something you'll regret. Thryduulf (talk) 17:00, 2 May 2026 (UTC)

You first. Gnomingstuff (talk) 04:22, 3 May 2026 (UTC)

adding my 2 cents:

1. It's fine to tag, but tags are also easily removable and if contested I would expected any attempts to re-add include some AISIGN in the article itself.

2. Per the guideline no, but if an editor is super 110% sure there is nothing wrong and are willing to risk their reputation on it, they can invoke WP:IAR (and I wouldn't support tagging if IAR'ed). If the content does turn out to violate policy, I would look at it the same as if someone was arguing to keep copyrighted or false content (or whatever policy ended up being violated). Jumpytoo Talk 01:38, 30 April 2026 (UTC)

1. Yes, tagging on "circumstantial evidence" like a broader pattern of an editor's LLM dependence is acceptable. If a tag was added on this basis, and then removed only because the article itself contained no obvious "tells", the tag should be re-added, because the check did not address the source of the problem. Our readers deserve to know when we are falling short of our ideals.

2. LLM-generated content without any other policy violations is not allowed in mainspace and should be tagged. Every example of it is a case where we have let something morally wrong happen on our platform. Stepwise Continuous Dysfunction (talk) 21:29, 2 May 2026 (UTC)

My understanding:

The rationale for WP:NOLLM states it is because it "often violates several of Wikipedia's core content policies", so removal requires that to be the case. From my point of view it is more important to keep AI slop off Wikipedia because the LLMs train on Wikipedia and this would set up an undesirable feedback loop but that is not what the guideline says. Over time, AI generated material should become increasingly indistinguishable from human generated content so we need to be alert to disruptive editors tagging anything they don't like as AI generated. Hawkeye7 (discuss) 20:55, 30 April 2026 (UTC)

I think the problem is that the current version of WP:NOLLM relies on LLMs breaking core content policies to justify its existence, but then goes on to prohibit LLM text in articles regardless of whether or not there is any evidence of that. Therefore, there's no clear answer to question 2. Personally, I would love to see an ideological ban of LLM-generated text across Wikipedia (and the Internet for that matter), but that is never going to happen for a number of reasons, not the least of which is that LLM text is becoming increasingly hard to distinguish from human text. In the long term, there is no practical way to completely prohibit users from adding LLM text to Wikipedia. LLMs are destroying completely changing society as we know it, and Wikipedia is not going to be exempt from that. The best we can do for now is to remove content that either a) clearly violates WP:V, WP:NOR, or WP:NPOV or b) has clear WP:AISIGNS, while those still exist, because it is likely to violate those policies, for as long as that’s actually true. I₂Overcome ^talk 04:40, 1 May 2026 (UTC)

In my view, WP:NOLLM ought to be taken to allow any editor to tag and/or remove any content that is AI-generated, without that having to be justified in terms of content in each individual case. If another editor thinks that the content should be untagged and/or kept, it should be open to that editor to disagree and argue that the content is not AI-generated (in which case the grounds for thinking that it is AI-generated can be debated on the talk page), or to check and rewrite the content personally on the basis of the sources given (in which case the content would effectively be written by a human editor, and any tags could be justifiably removed). But no-one should be able to argue that content is AI-generated but should nevertheless be kept since no content issues can be detected: such content may well have issues of sourcing etc. that might not be easily noticed by editors, and we should err on the side of excluding content that may not have been checked against the sources by a human editor. Dionysodorus (talk) 21:28, 2 May 2026 (UTC)

• 1: Yes, if there is reason to believe an article may incorporate text from a large language model then it is reasonable to add a tag stating that the article may incorporate text from a large language model.

2a: No, WP:NOLLM has no exceptions for "unless the edit is one of the good ones". Generated content is prohibited. 2b: No, see 2a. fifteen thousand two hundred twenty four (talk) 07:57, 3 May 2026 (UTC)

• May not be the best place for this but just wanted to point out a use of LLMs that seems legitimate to me. I recently rewrote the lede for Broadcast journalism with a helping hand from an LLM (though of course subsequently copyedited). It is clearly a dramatic improvement, and without the ability to use the tool I would have left it in its previous state. Point is in situations where a summary is needed, but no one has bothered to write one, LLMs provide a quick fix without any concern over sourcing (no need for references in lead). It's easy to then check the summary against the article for consistency. Is this considered an offense? Happy to self-revert if so. Wh1pla5h99 (talk) 16:14, 3 May 2026 (UTC)

  Is this considered an offense? – Yes, you have used an [LLM] to generate ... article content which is prohibited. Please self-revert.

  And no, summarizing is not a good application for LLMs as the model can still predict incorrectly and introduce errors and bias. The WMF experimented with it a while ago and the community nearly caught fire . fifteen thousand two hundred twenty four (talk) 16:37, 3 May 2026 (UTC)

  Got it, will revert! On errors, my assumption was that it would be possible to verify the text afterwards, as this is a rather straightforward task. (At least in this case it was.) Wh1pla5h99 (talk) 16:44, 3 May 2026 (UTC)

  In this case, I could easily find issues with the lead. It adds a source, , to the statement Broadcast journalism is the field of journalism that reports news and current affairs through electronic media, primarily radio and television, and increasingly through online video media., but online video media is not mentioned in the source (a listing for an instruction program about broadcast journalism), and is only discussed in the article as a concurrent force (and social media as a tool for stories) rather than as a medium of broadcast journalism.

  The risk when summarizing, especially with broad topic articles like this one, is that the model will mix up the content it has to summarize with knowledge from its internal weights, or extrapolate from either, leading to inferences not present in the article itself. Chaotic Enby (talk · contribs) 16:38, 3 May 2026 (UTC)

  I actually checked that source and added it manually, so that one can be chalked up to human error, although it wasn't supposed to perfectly support everything in the sentence as the rest of the article should do that. Wh1pla5h99 (talk) 16:43, 3 May 2026 (UTC)

• 1. No. You shouldn't place a tag at the top of an article unless you've spotted something in that article that you suspect is LLM-generated content. You should also be clearly stating which part of the article you suspect to be LLM-generated. Don't send other editors off on a wild goose chase trying to find where the content that made you add the tag is when it's located on a different page.

2a. Yes, if other editors feel like it. WP:NOLLM prohibits the using LLMs to generate article content and establishes that editors who do so can be sanctioned, but it doesn't say anything about what should be done with the LLM-generated content itself. This means that WP:NOLLM is violated when someone adds LLM-generated content, but said content being on the article does not in and of itself constitute a violation of WP:NOLLM. That said, I certainly think that editors are permitted to revert edits that add LLM-generated content as those edits constitute a guideline violation, but there is nothing saying they must do so. Editors may instead decide to edit/ensure said content is policy-compliant and keep it.

2b. The template needs to be updated. The first sentence is causing issues. The second and third sentences are in line with what the guideline actually says, but the first sentence says it is forbidden for articles to incorporate (have) LLM-content, which is misleading as WP:NOLLM only forbids users from incorporating (adding) LLM content into articles (2a). It also just doesn't rhyme at all with the second and thir dsentences; if it's forbidden for articles to have LLM content (which by extension means such content should be removed), then it is redundant to also state that LLM content that violates policies should be removed. It's saying "Remove X" and then "Remove Y" where Y is a subset of X (it should only be saying remove Y).

For the sake of argument, let's say that problem with the template was solved. In that case it would not be appropriate to place it down on articles confirmed to have LLM-generated content. It only makes sense to place it down if there is content suspected but not confirmed to be LLM content. If it's confirmed then the editor shouldn't be placing down a template for someone else to deal with; they should either be removing that content or editing/ensuring it to be policy-compliant.

For the record my personal opinion is that blatant LLM slop discredits the encyclopedia even if policy-compliant and that any such text should be rewritten by an editor if it is to be kept, but no community consensus has been established to that effect. ⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 23:32, 2 May 2026 (UTC)

TL;DR:

• 1. Don't put cleanup templates at the top of the article because of something you saw on another page; you're going to end up sending editors on wild goose chases after the suspicious content.
• 2a. You can revert edits adding LLM-generated content as those edits are violations of WP:NOLLM, but as it is the edits themselves that constitute violations of the guideline rather than the mere presence of LLM-generated content on the article you don't have to remove anything simply because it's LLM-generated. This does not mean you have to analyze every bit of LLM-generated content to see if it's policy-compliant before you remove it; it just means you have the choice to ensure it is policy-compliant instead of removing it.
• 2b. Template needs an update because it's misleading. Regardless of that, you shouldn't be placing it on an article for content that has been confirmed to be LLM-generated. If the content itself is violating PAGs you should be either removing or fixing it, and if it isn't then you shouldn't be putting down a cleanup template because there isn't anything to clean up.

⹃Maltazarian ^{ᚾparley${\displaystyle \lor }$investigateᛅ} 06:36, 3 May 2026 (UTC)

I need some eyeballs and some opinions here: Wikipedia does not need you Egezort (talk) 19:21, 30 April 2026 (UTC)

 You are invited to join the discussion at Template talk:Main § Main article link to non-English Wikipedia. This discussion needs more input. I'm not really sure where to advertise it, but it's sort of policy-adjacent so I'm trying here. If there is a better place please let me know Thryduulf (talk) 11:39, 3 May 2026 (UTC).

WP:NOTSTATS is frequently invoked as a component of our WP:NOT policy, in which it is part 3/4 of WP:INDISCRIMINATE. This especially applies to coverage of sport. NOTSTATS begins by ruling that Wikipedia articles should not be excessive listings of unexplained statistics. I don't think this goes far enough, and I propose removal of the word "unexplained". In addition, I propose addition of a stronger emphasis to the wording of WP:NLIST.

NOTSTATS then says Statistics that lack context or explanation can reduce readability and may be confusing. Provision of text is often cited as a reason to retain a set of statistics, but quite often the text only adds to the confusion, having been written with the obvious assumption that all readers understand the subject. INDISCRIMINATE rightly declares that Wikipedia is not an indiscriminate collection of information, and goes on to say that while editors may cite a database, Wikipedia itself should not be a database.

If a statistical table is a reasonable size, so that it fits comfortably within its main article, and if it is designed in a way that it readily provides information which would be difficult to render in narrative form, then there is no problem. For a typical example, see Australian cricket team in England in 1884, which is in WP:GA. This has three main tables, but they provide essential information, with statistics restricted to a necessary minimum (the match results). It would be impractical to impart any of that information in narrative form, so either a table or some form of list is necessary.

The problem is indiscriminate collection of statistics within excessive listings that look like something in a statistical database. NOTSTATS says Where statistics are so lengthy as to impede the readability of the article, the statistics can be split into a separate article, which takes us to WP:NLIST. This is, of course, a guideline and not a policy. But, it contains the key statement that a list topic is considered notable is if it has been discussed as a group or set by independent reliable sources.

Among statistical lists which have been discussed at AfD in recent times have been:

• Wikipedia:Articles for deletion/List of international cricket five-wicket hauls on Irish cricket grounds
• Wikipedia:Articles for deletion/List of international cricket centuries at the Sydney Cricket Ground
• Wikipedia:Articles for deletion/List of international cricket centuries at Kingsmead Cricket Ground
• Wikipedia:Articles for deletion/List of international cricket centuries at Carisbrook
• Wikipedia:Articles for deletion/List of international cricket centuries at Dubai International Cricket Stadium
• Wikipedia:Articles for deletion/List of international cricket centuries at the Gabba
• Wikipedia:Articles for deletion/List of international cricket centuries at the Adelaide Oval

In all of these, the main arguments for deletion were NOTSTATS and NLIST, the latter because there are never any independent, non-database sources which discuss the intersection as a group. Even so, given the weaknesses of both NOTSTATS and NLIST, there was considerable opposition to the nominations, and there have been cases of nomination failure because of those weaknesses.

While a particular achievement is important to individual players, and should of course be mentioned in their biographies, intersection between the achievement and, for example, the venue or the timespan, is not a notable topic unless there are sources which verify that it is notable. The intersection must be considered in terms of defining characteristics. Taking one of the above cases as an example, a batter scoring a century is not a defining characteristic of the Adelaide Oval as a venue. Neither is any one century at the Adelaide Oval a defining characteristic of centuries. Therefore, the intersection is not a notable topic unless reliable sources verify that it is (they don't).

NOTSTATS and NLIST only go so far towards resolving the issue of indiscriminate statistical lists. We need to apply NOTSTATS to any "excessive listing of statistics", whether there is context or not, and we need NLIST to emphasise, not merely mention, that "a list topic MAY ONLY be considered notable if it has been discussed as a group or set by independent reliable sources". With that level of certainty, the issue of indiscriminate lists will become clearcut, and considerable time will be saved at AfD. In addition, this proposal if implemented will go some way towards restricting overuse of statistical databases.

Many editors have shown interest in the statistical list AfDs, so I'm pinging those who I know to have made a contribution (for or against) to any of the above discussions, and to some similar ones:

• @11WB, Abhishek0831996, AirshipJungleman29, Ajf773, Ankurc.17, Aspirex, Bduke, BilledMammal, Clog Wolf, Doczilla, and Dronebogus:
• @Dz5t 8O12, Eddie891, Fram, Hyperbolick, Ktin, LibStar, Liz, NavjotSR, Nigej, OliveYouBean, Perry Middlemiss, Randykitty, and Reywas92:
• @Rockycape, Rugbyfan22, Servite et contribuere, Shellwood, Shrug02, SilkTork, Sirfurboy, Spartaz, Spiderone, Spike 'em, and StAnselm:
• @Star Mississippi, Stifle, Theleekycauldron, Turnagra, Vanamonde93, Vestrian24Bio, and Wakelamp:

If I've overlooked anyone, I apologise. Please feel free to invite anyone else who is known to have an interest. Thanks, Jack (talk) 19:08, 3 May 2026 (UTC)

I might be missing something, but I'm not sure where the issue lies or what exactly you're trying to fix - in those AfD examples you've mentioned, only one of them had what I'd call "considerable opposition", and all of them resulted in the nominated pages being deleted. It seems like the current level of PAG is working? Turnagra (talk) 19:31, 3 May 2026 (UTC)