Wikipedia talk:New pages patrol/Reviewers

Please consider joining us for the May 2026 NPP backlog drive.
Sign up today.

TutorialDiscussionNew page feed

• Sorted
• NPP Browser

• Easy
• Reports

Reviewers

• Resources
• School
• Top

Curation tool
SuggestionsCoordination

• Talk
• Awards
• Newsletter

• May 2026 backlog drive

NPP backlog

Articles

14570 ↓214

Oldest article

2 years old

Redirects

34203

Oldest redirect

6 months old

Article reviews

2514

Redirect reviews

2906

• There is a very large articles backlog
• There is a very large redirects backlog

NPP backlog

NPP unreviewed article statistics as of May 18, 2026

Am I crazy or are the majority of new articles LLM-assisted?

Articles by newish editors are too perfect nowadays. Here are some examples: Theo Angelis, Golden Landmark Shopping Complex, Como Beach, Euposaurus. I always have typos when I write long articles but most new articles never have any. Their references are always formatted perfectly as well. I haven't seen a bare URL in forever. To prove my point about the aforementioned Theo Angelis article, I went to a similar Washington state judge article: Oscar Raymond Holcomb (created in 2017). I found a typo pretty quickly: "where he lived for the next for 15 years". I also went to the judge right before Angeles in the succession box, Barbara Madsen, and found bare URLs. ~WikiOriginal-9~ (talk) 20:07, 6 April 2026 (UTC)

Yeah, it's depressing. Makes it really hard to check the articles. I've been finding that quite a few of them are so close paraphrases of the cited sources that they should probably be reported as copyvio, which is a huge pain to do. I just AfD'ed this HUGE bundle of AI-generated articles: Wikipedia:Articles for deletion/Traditionsschifferschein. And here is a discussion about whether close paraphrasing should be an additional criteria for G15: Wikipedia talk:WikiProject AI Cleanup#Add criteria to G15: Close paraphrase of source. Lijil (talk) 20:41, 6 April 2026 (UTC)

It's sad. I think Claude is getting much better with writing articles, whereas ChatGPT was very easy to spot. It's also very sad because (for example) Theo Angelis clearly meets notability guidelines (I was actually looking into his 2026 re-election campaign just a few days ago and was surprised not to see an article), and the page has no glaring errors, but Wikipedia losing the human touch. aesurias (ping me in your reply, or I won't see it) (talk) 22:05, 6 April 2026 (UTC)

As if on cue, I get this stinking pile of robo-faeces.Boynamedsue (talk) 08:14, 7 April 2026 (UTC)

Yikes, that's awful. Are new editors getting clear notifications that AI-generated content is not accepted? Or do they know and they just don't care? Lijil (talk) 12:00, 7 April 2026 (UTC)

No, and yes. Well or they insist that their LLM is good and doesn't have hallucinations, and insist we check every single line they ever wrote. I've been doing some RCP AI spot checks and the rate appears to average out at between 3 to 6 obvious (like, blindingly obvious) AI edits per hour. Fermiboson (talk) 12:06, 7 April 2026 (UTC)

I'm not sure if they are getting guidance or not. A lot of people have banner blindness anyway. Every BLP has a big red banner in the edit window, but most people don't even notice it (myself included). ~WikiOriginal-9~ (talk) 13:04, 7 April 2026 (UTC)

Are new editors getting clear notifications that AI-generated content is not accepted...

A question related to this.

First, some context: T420258 describes an LLM-specific version of Paste Check. The idea being when we detect people pasting content from an LLM into the viusal editor, we'd surface an Edit Check that would 1) inform people of Wikipedia:Writing articles with large language models, 2) require people to explicitly indicate whether they did or did not use an LLM to generate the content they're pasting, and 3) in the background, tag edits in which this Check was activated.

A resulting question: how (if at all) can you see an intervention of this sort being helpful? PPelberg (WMF) (talk) 23:02, 12 April 2026 (UTC)

This would be super helpful! We know from the other tests that the intrusion of edit check makes a measurable impact. Even if it can only detect half of the big LLMs, it should make an impact. —Femke 🐦 (talk) 06:47, 10 May 2026 (UTC)

@PPelberg (WMF), the sooner this comes out the better. -- asilvering (talk) 21:45, 10 May 2026 (UTC)

@Femke + @Asilvering: noted! You can expect me to be back in touch about this in the next ~2 weeks. The Editing Team is exploring what an initial proof of concept could look like and the level of effort involved to implement it. PPelberg (WMF) (talk) 22:09, 11 May 2026 (UTC)

@Femke+ @Asilvering: Okay. I'm back with an update: an initial version of an LLM-specific iteration of Paste Check is available for you all to try. See testing instructions below.

Before that: while we are eager to hear any and all questions/ideas this exploration brings up for you, we are especially interested in learning:

1. What options do you think should appear in the survey that appears within the Check when people elect to Keep the text they're pasting?
2. What en.wiki policy/guideline do you think should be linked within this Check? E.g. WP:G15?
3. What facet(s) of the Check do you think would need to be configurable on a per wiki basis? E.g. Account state(s) the Check should apply to, sections the Check should not activate within, whether the Check should activate on quoted content, minimum amount of characters someone needs to have pasted for the Check to show, etc.

Last thing, if there are pages/people who you think would be excited to try this out/offer feedback, I hope you will let me know and/or share it with them directly!

How to try LLM Paste Check

1. Visit https://564a50573d.catalyst.wmcloud.org/wiki/Regent's_Park?veaction=edit&ecenable=1 on desktop or mobile
2. Copy at least one full paragraph of text from the web interface of ChatGPT, Claude or Gemini
3. Paste the text you copied in "2." into the edit session you started in "1."
4. ✅ Notice the Potential AI-generated content Check appear

PPelberg (WMF) (talk) 21:55, 14 May 2026 (UTC)

Absolutely fantastic that you're making fast progress, as AI slop is really starting to overwhelm patrollers. (1) I think these should maybe align with the two exceptions in WP:LLM: That is, "This is a manually checked translation", "I have verified this is an accurate copyedit" and "None of the above". (2) In terms of the guideline to cite, WP:LLM is better than G15. G15 is only about the worst of the worst. (3) those configuration options seem sensible. In practice, I think we would configure which people this applies to and which sections this applies to (all of them?). —Femke 🐦 (talk) 07:27, 15 May 2026 (UTC)

For articles whose subjects are clearly notable, an alternative to draftifying is always to stubify. ​I feel like that option isn't given adequate consideration in guidelines like WP:NPP. lp0 on fire () 12:07, 7 April 2026 (UTC)

How would you go about doing that? I can foresee the edit getting reverted for removing sourced info. Orange sticker (talk) 13:00, 7 April 2026 (UTC)

Delete the entire slop and replace with one liner and a hatnote to expand content that isn't LLM generated ~ 🦝 Shushugah (he/him • talk) 13:15, 7 April 2026 (UTC)

No need for the hatnote; just temove it outright. As for not getting reverted, "removal of sourced content" isn't a reason to revert, and if the alternative is outright deletion there's nothing wrong with turning it into a microstub. lp0 on fire () 13:30, 7 April 2026 (UTC)

It's a reason I've often seen given when reverting vandalism or similar. I guess if "LLM generated" was added to the text at WP:STUBBIFY we could link to that in the edit summary as explanation. Orange sticker (talk) 13:43, 7 April 2026 (UTC)

Yeah, if a recent changes patroller is reverting valid stubifications by new page reviewers on the grounds that they removed sourced content, they probably shouldn't be patrolling recent changes. lp0 on fire () 14:16, 7 April 2026 (UTC)

I accidentally clicked "send" without finishing this, but oh well. lp0 on fire () 16:41, 7 April 2026 (UTC)

I just stubified an article that seems notable but the information wasn't supported by the inline citations, along with other signs of LLM. What now? Mark as reviewed? Still seems unsatisfactory: Disques Flèche. Orange sticker (talk) 08:50, 8 April 2026 (UTC)

Re your earlier "It's a reason I've often seen given when reverting vandalism", the reason I usually give is that I'm reverting an unexplained removal of sourced information. If the explanation that it is removal of AI slop were given in the edit summary, I would not revert it. —David Eppstein (talk) 19:34, 8 April 2026 (UTC)

I wonder if a caution about LLM should be added to the standard Welcome texts? Ldm1954 (talk) 12:09, 7 April 2026 (UTC)

We need more than that. When pressing submit on a new article, the creator should be taken to a page where they tick a box stating "I have not used a Large Language Model in the creation of this article. I accept that if evidence emerges that this article includes text from a Large Language Model, the article will be deleted and I may be banned from English Wikipedia."Boynamedsue (talk) 16:48, 7 April 2026 (UTC)

We could do that, although that would require a wider concensus and it sounds a bit aggressive to me.

I have started a discussion on adding LLM the welcome templates at the welcome talk page, please feel free to chip in there. Ldm1954 (talk) 17:08, 7 April 2026 (UTC)

I wonder if we can modify the local lp0 on fire () 17:37, 7 April 2026 (UTC)

I love this idea. Maybe it only happens the first X times someone edits. Lijil (talk) 19:49, 8 April 2026 (UTC)

Sorry, I didn't realise how bad my internet was. What I meant to say here was I wonder if we could modify the local message for mw:Edit check/Paste Check to include an AI warning. lp0 on fire () 20:39, 8 April 2026 (UTC)

It seems to have already been done: see MediaWiki:Editcheck-copyvio-description. OutsideNormality (talk) 03:32, 10 April 2026 (UTC)

A checkbox was brought up during one of the RfCs on AI policy, and there was some opposition to it because some editors thought it would encourage people to use AI or give the impression it was permitted. Of course things have changed since then so it may be worth revisiting. Gnomingstuff (talk) 13:25, 10 April 2026 (UTC)

There needs to be a coordinated approach across sections. I have just draftified the last 3 creations of the user who pressed send on AI's article Draft:Ideology of the Confederate States of America. This is probably about the 30th I've done for this reason in the three weeks I've been an NPP. There's an avalanche of this crap. As of now, do we have a local consensus in favour of stubifying all AI-generated articles? And what happens when the users kick off?Boynamedsue (talk) 16:15, 7 April 2026 (UTC)

Do you mean stubbifying just the ones where the subject is likely notable? I'd be interested in applying that approach and seeing how it goes. Orange sticker (talk) 16:20, 7 April 2026 (UTC)

There is very strong consensus against writing with LLMs, which implicitly equates to consensus in favour of removing that content. If the topic isn't obviously notable, it's usually uleasi lp0 on fire () 16:36, 7 April 2026 (UTC)

Sound. That makes sense.Boynamedsue (talk) 16:38, 7 April 2026 (UTC)

I stubbify when the subject is clearly notable as demonstrated by the references. The exception - regardless of notability - is prolific AI contributors: editors repeatedly pushing out slop despite warnings. In that case, I'm concerned they'll just restore the content, and I don't want to end up feeling obligated to babysit the page. For new users, though, I've had pretty good luck with stubbifying. I try to politely explain the AI signs I'm seeing and frame it around our shared goal of marking the article as patrolled for mainspace. The handful of times I've done this it's gone well. Zzz plant (talk) 16:39, 7 April 2026 (UTC)

In case of prolific AI editors who will just restore the content, it's probably better just to take them to ANI (or just ping {{@AINBA}} at WP:AINB) and keep the stub. Even if you only keep a line, removing content because of what someone might do to it doesn't feel right; even in an extreme case protection would still be preferable to outright removal. lp0 on fire () 16:45, 7 April 2026 (UTC)

Draftify may be safer than WP:TNT, because then the article doesn't get marked as reviewed. TNT is also valid, but reviewers should probably also watchlist it to make sure the slop isn't put back as soon as you mark the article as reviewed.

Note that when someone undo's one of your edits, the only reason you get a ping is that your username is linked in the default edit summary. If that edit summary is changed, or the person just copy pastes the old revision text back in, you will get no notification.

Above it is mentioned that some recent changes patrollers revert TNTs. That seems odd. Make sure to use a clear edit summary to help with this. Something like: "WP:TNT this article due to WP:LLM use", or even simply "WP:NOLLM". –Novem Linguae (talk) 20:20, 7 April 2026 (UTC)

We also shouldn't limit this to new users. I have seen countless inactive accounts come back to make these AI articles. – The Grid (talk) 20:42, 7 April 2026 (UTC)

There's a kernel of truth to this, though, in that one of the main vectors of AI-generated content and rewrites is Newcomer Tasks. Gnomingstuff (talk) 20:03, 8 April 2026 (UTC)

If you're sure about this, then perhaps we should look into adding some "no AI" warning text to the Newcomer Tasks interface. Can you provide a URL to a newcomer task? I'll use it to poke around and see how easy/hard it would be to do this. (Might be as easy as editing a page in the MediaWiki namespace.) –Novem Linguae (talk) 03:13, 10 April 2026 (UTC)

Believe me, we've tried, unfortunately this appears to require WMF to do it, and, well.... Gnomingstuff (talk) 13:23, 10 April 2026 (UTC)

I would not be surprised if the percentage was high -- I keep thinking about that New York Magazine article that claimed over 90% of students had already used ChatGPT for stuff by January 2023 (i.e., only a few months after its release).

I don't have an actual sense of proportion since I don't do new page patrolling (no bandwidth or interest, sorry), and it's hard to get an accurate sense of the proportion of older edits since things get deleted/reverted. But based on what I have seen, this has been an issue since early 2023, and really started to snowball in mid-2024 (when GPT-4o came out), and by 2025 was an avalanche. In other words, there was just as much AI stuff a year ago, it just was not being detected despite being obvious in retrospect.

I don't think newer AI output is undetectable, especially if you've seen a lot of it. (Based on these examples newer Claude models are the only ones that sound that much different from other LLMs, including GPT-5. However, I suspect most people use GPT-5.) But it's harder to explain to people why it's detectable, in the same way that it's hard to explain anything where you've seen 10,000 examples and they've only seen one. Gnomingstuff (talk) 20:00, 8 April 2026 (UTC)

I have been (am currently) doing RC monitoring for obvious AI edit summaries in order to get a sense of the rate at which obvious, lazy copy-paste AI edits are coming in. The rate currently varies between one per 3 minutes to one per 15. Fermiboson (talk) 20:09, 8 April 2026 (UTC)

extremely believable, unfortunately Gnomingstuff (talk) 20:16, 8 April 2026 (UTC)

It's possible with VisualEditor to create rich references easily nowadays, so a new editor being able to create references perfectly is not unusual at all (you can tell other than looking at the Tags is that the ref names are usually ref name ":0"> or similar). Also some editors like to draft in their sandbox and then copy and paste into the mainspace, it's what happened for the Golden Landmark Shopping Complex and Euposaurus articles, so in those cases a large creation edit is not unusual. Raw LLM output is still very detectable, to confirm this I asked Sonnet 4.6 to generate me a Wikipedia article, and the output is immediately obvious that it is LLM generated. Finally, instead of looking at typos as a sign of human editing, I'd look at things that the average Wikipedia article has, for example categories, short description, or an infobox. If the editor didn't add those and the article otherwise looks good that could suggest human writing, as LLM knows very well the average format and will almost always include those if you ask it to generate an article (even if it does so incorrectly like with DOA categories). Jumpytoo Talk 01:59, 9 April 2026 (UTC)

I support default stubbing.★Trekker (talk) 11:21, 11 April 2026 (UTC)

Agree. If there's a potentially valid (WP:N) article, it makes little sense to ostensibly banish the subject because of the taint of LLM. It seems more productive to strip out the LLM-chaff and keep a clean stub (e.g., I'm not likely to go back and re-create this, but in my view, as the subject appears notable and is mentioned elsewhere in our pages, it would have been worth preserving as a stub). -- Cl3phact0 (talk) 11:59, 11 April 2026 (UTC)

Infobox use policy

Via a page on my watchlist which I monitored after doing some WP:NPP, I came across mass changes to page infoboxes that I question, and would like to check here on policy.

~2026-36481-5 (talk · contribs · deleted contribs · logs · filter log · block user · block log) has been rapidly removing large numbers of "Infobox needed" and "Photograph needed" maintenance tags from talk pages. In some cases an infobox was present, so they are OK edits. In many cases there was no photo, so the edit seems to be dubious; no talk concensus per MOS:INFOBOXUSE. In some other cases the editor used the edit description "Short article" which I do not see as a justification. (In many others there was no edit description.)

I have tagged on their talk page twice, no explanation from them as yet. Are these removals some policy, or should I bounce this to WP:ANI; checking all these edits as some seem to be inappropriate is a lot of work for volunteers.

N.B., I have also asked at WT:MOSIBX, but it is perhaps off-topic there. Ldm1954 (talk) 11:32, 13 April 2026 (UTC)

Speaking as an editor who has spent time this year clearing and reclassifying a number of photo requests, there are a few where I've concluded there is no possible way to get a photo that would be relevant (example) or when I've BLARed an article (example) and thus removed the request. Would be different if it's happening on a mass scale and with situations where a photo could be possible. Dclemens1971 (talk) 16:27, 13 April 2026 (UTC)

How to become New page reviewer

i want to become NPR.. Mexico's Claudia (talk) 11:39, 13 April 2026 (UTC)

@Mexico's Claudia Please review the requirements at WP:NPRCRITERIA. Applications are made at WP:PERM, but please do not apply now, as you are a brand-new editor and you do not currently meet the requirements for this permission. Spend time engaging constructively by improving articles, writing new ones that comply with our guidelines, and/or participating in deletion discussions, then you can apply once you meet the criteria and have more experience. Good luck. Dclemens1971 (talk) 14:35, 13 April 2026 (UTC)

Thank you for explaining (: Mexico's Claudia (talk) 16:23, 13 April 2026 (UTC)

another question if a draft is weak under AfC submission can we leave a review comment instead of accepting it? Mexico's Claudia (talk) 19:08, 13 April 2026 (UTC)

@Mexico's Claudia You need to be an AfC reviewer to review/comment on drafts. That's a separate permission, and you shouldn't be reviewing drafts without it. HurricaneZeta^C 19:11, 13 April 2026 (UTC)

The requirements to be an AfC reviewer are the same as to be a new page reviewer, and it must be applied for at WP:AFCP. Please focus on improving and creating content and revisit NPR/AFC work when you meet the criteria. Dclemens1971 (talk) 19:13, 13 April 2026 (UTC)

👍Alright @Dclemens1971@HurricaneZeta Mexico's Claudia (talk) 19:16, 13 April 2026 (UTC)

OP has been sock-blocked. ClaudineChionh (she/her · talk · email · global) 23:30, 13 April 2026 (UTC)

I had some kind of gut feeling here - strange Squawk7700 (talk) 16:29, 14 April 2026 (UTC)

Automatically patrolling redirects created from AfD closes

@Vanamonde93 had the idea that we could automatically patrol redirects created as a result of AfD closes as merge or redirect. I'm wondering if this could be accomplished using @DreamRimmer's bot, which already does a bunch of automatic redirect patrolling. My rough idea of the conditions would be that the bot checks the AfD page (e.g. whether it says "the result was redirect") and check whether the editor doing the BLAR is extended confirmed (as non-xcon users are generally not qualified to close AfDs, and to prevent abuse). Maybe creating a {{R from AfD}} would be helpful. But I have no idea how to program a bot and what's feasible. Bringing this here for broader input. Toadspike [Talk] 19:40, 15 April 2026 (UTC)

Seems like a decent idea. I would Support. Kingsmasher678 (talk) 22:41, 15 April 2026 (UTC)

Support, every little bit helps! SnowyRiver28 (talk) 06:43, 16 April 2026 (UTC)

Happy to implement this. Let's wait a few more days for more opinions. @Novem, what do you think? – DreamRimmer ■ 15:42, 16 April 2026 (UTC)

Sure. Any reasonable addition to the redirect autopatrol bot algorithm sounds fine to me. Lots of people responded positively here so should have a good consensus to link to in the BRFA. –Novem Linguae (talk) 05:53, 17 April 2026 (UTC)

Seems fine to me _signed, Rosguill ^talk 16:15, 16 April 2026 (UTC)

That also makes sense to me. Aszx5000 (talk) 17:04, 16 April 2026 (UTC)

Yes please. --Northernhenge (talk) 18:06, 16 April 2026 (UTC)

I honestly thought DR had this as part of the bot already. Vanderwaalforces (talk) 18:41, 16 April 2026 (UTC)

yes, do it--Ozzie10aaaa (talk) 01:20, 17 April 2026 (UTC)

Sounds like a good idea to me. Vestrian24Bio 03:02, 17 April 2026 (UTC)

As long as it is reasonably prevented from being able to be abused, that sounds like a really good idea. Squawk7700 (talk) 12:10, 21 April 2026 (UTC)

Review and Redirects Draft

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

Review this Dil Sher Brohi a new Article Page. and this Draft page Draft:Dil Sher Brohi to redirect existing page Dil Sher Brohi. Thank you. Kashif123m (talk) 18:19, 16 April 2026 (UTC)

@Kashif123m:  Done You can request G7 deletion of Draft:Dil Sher Brohi 2 and let the AfC process handle Draft:Dil Sher Brohi which you have currently submitted for review. Please do not create again the same draft duplicates in article main space. CONFUSED SPIRIT(Thilio).Talk 19:49, 16 April 2026 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Ancient Refspam

Earlier today I came across a number of edits by User:Stevebillings in 2013, see this edit list, many of which are to a book by Stephen Billings. This looks like a fairly obvious case of Refspam, as pointed out in 2015 on Reddit. I reverted one, I wonder about reverting the others. My caveat is both the age and that Stephen Billings passed away in 2022. Thoughts? Ldm1954 (talk) 06:01, 18 April 2026 (UTC)

We always need to cite the most reliable independent sources that give an authoritative account of the subject. The date that the sources were added to the article isn't in itself relevant, so the age of the edits isn't an issue. If they're still the best sources, they should be there. If there are now better ones, they should be there as well or instead. WP:REFSPAM says "Often [spam links] are added not to verify article content" which would be a significant problem but I don’t know if that's true here. We often revert sockpuppet edits (and perhaps topicbanned edits) however valid they are, but otherwise it would be unusual to revert a good edit to the best source just because we doubt the motive of the editor. It's sad to know that he’s passed away, but that knowledge shouldn't influence us. There must be a great many we don't know about. Northernhenge (talk) 20:08, 18 April 2026 (UTC)

Page mover tool

As an NPP reviewer, I do come across situations where I need to make an uncontroversial page move and the ability to not leave a redirect and/or the ability to do round-robin page move, would be very helpful. Is this a valid reason to ask for the page mover right, and it is understood that this is a natural consequence of doing a lot of NPP work? thanks. Aszx5000 (talk) 12:33, 18 April 2026 (UTC)

Yes. It's why I got PM myself. -- asilvering (talk) 12:35, 18 April 2026 (UTC)

Okay great - thanks for that. Aszx5000 (talk) 16:07, 18 April 2026 (UTC)

I applied but failed unfortunately. I actually think the PM tool should come with the NPP tool as I must have over 50 "orphan" redirects but it failed to convince them at WP:RFP/PM. Aszx5000 (talk) 09:45, 19 April 2026 (UTC)

Even though NPP work does sometimes feel like creating clean-up work for other people, it's good that if I propose anything destructive, there is always a second pair of eyes on it from someone with the appropriate user rights. --Northernhenge (talk) 16:57, 19 April 2026 (UTC)

I focus most of my time of estimated B grade (and above) new articles, so there is a lot of clean up in places, but I do find that satisfying and it is appreciated by the editors.

I think you make a good point, and in fairness to those at PM, they were happy(ier) with me sending in more requests to RM (whether routine or otherwise), which I am going to do and see how I get on.

Thanks for all your attention ! Aszx5000 (talk) 08:54, 20 April 2026 (UTC)

Second Opinion Regarding Possibly Promotional Page

Hi!

I was going through the new page list and came across the article Dart Foundation, which seems to be extremely promotional. I wasn't sure whether to nominate the page for deletion or tag it with an advertisemennt/promotional content mark. Does anyone have thoughts on how to go about this? Any help would be appreciated. Thanks! Losipov (talk) 15:43, 23 April 2026 (UTC)

Definitely promotional in tone and offering excessive/non-encyclopedic details, but I think the subject is notable. Just tag it accordingly but I think you can mark it as reviewed. Dclemens1971 (talk) 15:54, 23 April 2026 (UTC)

Have to disagree there. Notability is not the sole criteria for marking as reviewed. The article meets the criteria of WP:G11, including would need to be fundamentally rewritten to serve as encyclopedia articles, and possibly WP:G15. I'd leave the article unreviewed and tag for G11 and/or G15. John B123 (talk) 20:52, 23 April 2026 (UTC)

I don't think it reads particularly as LLM-generated, just badly written and promotional. Editors have been doing unnecessary bolding and bullet points since long before ChatGPT was released. Dclemens1971 (talk) 00:31, 24 April 2026 (UTC)

That may be true, but it’s very common now with LLMs. See WP:AILIST Plus, the article has been edited to remove some more obvious signs, like the sentence: "These grants reflect the Foundation’s broader commitment to strengthening community services and advancing public interest programs." I₂Overcome ^talk 01:22, 24 April 2026 (UTC)

I think we do find people copying AI-style in non-AI text now, annoyingly.--Boynamedsue (talk) 12:35, 25 April 2026 (UTC)

It’s also LLM-generated (see WP:AILIST). When the subject is notable, some editors prefer to just tag it and let someone more experienced deal with it later. Others will reduce the article to a stub, removing all the LLM-generated, promotional text. Some editors nominate for deletion, since the whole article is in violation of WP:NOLLM. If the article is new enough (this one isn’t), you can draftify it. Any of these are acceptable, since there is currently no consensus on the best way to handle LLM-generated articles. I₂Overcome ^talk 20:23, 23 April 2026 (UTC)

How far is too far with AI tagging?

I want to raise an issue for discussion that I suspect many NPP reviewers have come across -- when is it right and wrong to tag for AI. Please do not get me wrong; unreviewed LLM with hallucinations is awful. However, should we tag anything and everthing if there is an indication and/or the editor has used LLM in the past? My opinion is that we should only tag when there is clear evidence of inappropriate information being added, and/or talk page essaying.

I have seen some tagging in the past which I felt was too abrupt, a very recent case is Gareth Thomas (materials scientist) where Gnomingstuff added an AI tag with The reason given is: user has demonstrated a pattern of AI edits. I had just finished NPP review of the page, and based both upon the sources there and (as it happens) my personal knowledge since I work in the field there were no issues. I have reverted the tag twice now as inappropriate.

It is true that User talk:LocusAndLeaf was warned (level 1) by AntiDionysius in January, and Gnomingstuff gave a non-template warning, but I do not think that means that all their articles should be tagged without detailed indicators. Since we are the people that do a lot of tagging, some discussion may be good to setup some policy of when to warn/tag or not. Ldm1954 (talk) 01:48, 28 April 2026 (UTC)

Note that this user has scolded me about the G15 speedy deletion template and when you can and cannot use it, when I did not actually tag the article for speedy deletion. So right off the bat, they either are deliberately lying or do not know what our AI guidelines are.

So, to refresh on what those guidelines are:

1) The text of WP:NEWLLM: [...] the use of LLMs to generate or rewrite article content is prohibited, except for basic copyedits and translation of material from other language Wikipedias. The guideline does not say "the use of LLMs to generate or rewrite article content is prohibited, unless someone at NPP thought it was OK." If you don't like that, then start another RfC to get the guideline changed, and good luck.

2) The AI-generated tag is not a speedy deletion tag. It isn't even a regular deletion tag. It isn't even a "this is 100% unquestionably without a doubt AI and I am so confident about this that I personally dare God to torment me forever in hell right now if I am wrong" tag -- there's a certain=y parameter for those cases.

3) If someone is warned (with grounding) about LLM use, does not acknowledge that warning, and continues to edit in the same way, then yes, it is likely they're still using AI. Here is a recent talk page comment by the user in question. It is blatantly AI-generated, as are their other talk page comments. The same goes for their articles, which are noticeable enough that multiple people have independently pointed them out as AI. Gnomingstuff (talk) 04:09, 28 April 2026 (UTC)

@Gnomingstuff, right off the bat, they either are deliberately lying is an incredible assumption of bad faith, and if dealing with AI cleanup has gotten you this burned out, I think it's probably time to step back from it for a while. -- asilvering (talk) 04:23, 28 April 2026 (UTC)

The word "either" has a meaning. I didn't put a G15 tag on the article. They're implying that I did. There's no gray area here, either you tag something for speedy deletion or you don't, and there's also no ambiguity because edit history is public and anyone can go see for themselves that I didn't. So, logically, this means that the possible scenarios here are that they are lying, that they didn't bother to check whether their assumption was even correct (very ironic), or that they don't know the difference between speedy deletion or cleanup tags. I don't see how this is assuming bad faith, but accusing me of doing something that I literally, factually, verifiably-by-anybody did not do is totally OK. Gnomingstuff (talk) 04:50, 28 April 2026 (UTC)

Also, don't ping me. Gnomingstuff (talk) 05:01, 28 April 2026 (UTC)

The comment in question was here . My guess is that Ldm was making a rhetorical point, that G15 only forbids LLM text without human review, and thus reviewed text is allowed. That is not quite accurate, as NEWLLM has a much broader prohibition. I do not, however, find it at all fair to accuse them of lying. Toadspike [Talk] 06:25, 28 April 2026 (UTC)

And I don't find it at all fair to be scolded about G15 when I didn't put a G15 tag on an article (and almost never do). Rules for thee but not for me, I guess. Gnomingstuff (talk) 21:11, 28 April 2026 (UTC)

That was also my guess, but I have to admit that the point of bringing up G15 in this context was unclear to me. G15 doesn't matter; the general prohibition does. Stepwise Continuous Dysfunction (talk) 23:26, 28 April 2026 (UTC)

Firstly, Gnomingstuff (look no ping), you should have been here long enough to know that you can't say people are lying without really, really good reason and evidence. You will wind up at ANI if you do that. I would strike and apologise.

Secondly, it is absolutely fine to tag articles as possibly AI-generated where they were created by a user who hasn't responded to a tag or enquiry about AI-use on their talkpage. This is assuming the first detection of AI occurred in good faith and nobody else has questioned it. We all know ignoring messages is part of one kind of disruptive user's cycle of behaviour, and make no mistake, using AI to create articles is disruptive behaviour.Boynamedsue (talk) 06:11, 28 April 2026 (UTC)

Once again, I did not say that they were lying. The word "either" has a meaning. If I wanted to say that they were lying, I would just say that and not hedge.

What I was trying to say, and thought I was more than clear in saying, is that I didn't tag the article for deletion -- speedy or otherwise -- and yet they scolded me about the speedy deletion policy anyway. And I don't see anyone asking them to apologize for that. Gnomingstuff (talk) 21:29, 28 April 2026 (UTC)

Can we all just pump the breaks for a second? I'm not saying @Gnomingstuff handled this the right way. But I would like to add that AI cleanup is not easy work. You have people lie and break promises about their AI use a lot and it causes real burnout. I've had to check myself a few times while dealing with folks and AI cleanup so I know that the toll is real. I am hopeful that we can move past this and give each other the grace we deserve. Dr vulpes (Talk) 06:42, 28 April 2026 (UTC)

Objectively G is obviously correct in their actions, or at least acceptable, but going in two-footed like that is not ok. I'm not going to grass them up, but it is in their interests that somebody tells them not to act in this way or they might wind up getting topic banned.Boynamedsue (talk) 06:50, 28 April 2026 (UTC)

A question: is it right to repeatedly add the LLM tag when an independent reviewer (me) states that there is zero evidence of any hallucinations etc? I have no problem with adding a tag once to get cleanup, but insisting that the page is tagged forever just seems to be inappropriate. Ldm1954 (talk) 13:00, 28 April 2026 (UTC)

@Ldm1954, no, it is not. That's edit-warring. -- asilvering (talk) 16:15, 28 April 2026 (UTC)

@Ldm1954: Even ignoring the fact that it is edit warring, it seems like a counterproductive waste of time. Honestly, my hot take is even if it is generated by an LLM, if its a literally perfect article, IAR should apply. To me, actioning an article in that case seems like process for the sake of process. Sure, 99% of LLM use onwiki is probably creating shitty promotional content, which is why I'm a fan of the anti-LLM policy, but to spend more than two seconds on the rare unnoticeable use case seems odd. ULPS ^{(talk • contribs)} 18:52, 28 April 2026 (UTC)

I think the directions for the Template:COI#The article should have a specific problem tag apply here and to pretty much every maintenance tag: if there's nothing in the article to fix, the tag should be removed, regardless of its provenance. –Novem Linguae (talk) 22:04, 28 April 2026 (UTC)

Except that the template isn't a "there are hallucinations" tag. It is a "this article may contain AI-generated content" tag. You can verify this for yourself by reading it.

The disconnect here, I guess, I am trying to argue from a standpoint of reality, which is public and visible and verifiable by anybody here, and everyone else apparently gets to just say shit regardless of whether it's true. Gnomingstuff (talk) 21:14, 28 April 2026 (UTC)

If the tag was originally placed because (for example) the page creator had a well-established history of creating pages with LLM, rather than because specific "hallucinations" were identified, then I don't think that checking for hallucinations is enough to give the article the all-clear. Stepwise Continuous Dysfunction (talk) 22:48, 28 April 2026 (UTC)

The tag was originally placed because the editor felt that the page creator was using LLM. I am fine with that. However, if the page has been reviewed and checked for hallucinations, OR, SYNTH and any WP:NOT then I cannot see a justification for reinstating the tag. Put even stronger, if several editors say "this article does not contain AI-generated content" then the tag should not be present. Ldm1954 (talk) 23:12, 28 April 2026 (UTC)

I don't see how any amount of checking the text of an individual article can alleviate concerns based on the page creator's behavior elsewhere, rather than the text of that individual article. It's just not addressing the root of the problem. If the issue with the article is X, and several editors say "this article does not contain Y", then the article should remain tagged for X. Stepwise Continuous Dysfunction (talk) 23:19, 28 April 2026 (UTC)

I will disagree with you, and use COI as an analogue (as somewhere else did somewhere in these discussions).

If there is suspicion of COI, then the page gets tagged with an explanation of the evidence provided. If the person denies the COI and there is no suspicion that they are lying, the tag is removed. If there is definite COI then we check for bias etc, and add {{connectedcontributor}} to the talk page, and require them to request edits (if the page is in main). Once the page has been checked for bias etc, and any issues cleaned up the COI tag on the main page is removed. If the page is 99% COI, e.g. un-notable self-promo, then it goes to CSD/PROD/AfD. (We have similar CSD/PROD/AfD for unrepairable LLM.)

If they do not behave on COI, then you warn them on their user page. If they continue then you take it to ANI. Similarly we warn then take to ANI for LLM, as most of us have done.

The page creator does not own either a COI page or a LLM page. Just as tagging a page in main as COI forever is inappropriate, tagging for LLM when all issues have been removed is not appropriate. Ldm1954 (talk) 23:41, 28 April 2026 (UTC)

In your analogy, you say, "If the person denies the COI and there is no suspicion that they are lying...". Here, I don't think there has been a denial of LLM use. (The user has been asked directly at least once but has not responded.) So, following your analogy, we are not yet at the point where the tag should be removed. Stepwise Continuous Dysfunction (talk) 00:20, 29 April 2026 (UTC)

No, if there is no response we are in the second case where we check for bias etc and cleanup.

N.B., the editor responded to my warning and admitted use of LLM in a different edit. So we are in the standard second case scenario.

N.B. I forgot a third case when the COI editor reverts cleanup. Then we escalate warnings to levels/3/4/ANI. Ldm1954 (talk) 00:35, 29 April 2026 (UTC)

Sorry, I meant LLM use regarding the specific article Gareth Thomas (materials scientist) about which the page creator has said nothing. Indeed, their reply to you is the first acknowledgment about LLM editing that I can find them giving. This seems too soon to declare that they have or have not denied using an LLM on the Gareth Thomas page. In your analogy, we are not yet at the point where we could tell between the first- and second-case scenarios. (I'm also not convinced that COI edits are the best comparison to make for LLM pollution.) Stepwise Continuous Dysfunction (talk) 00:44, 29 April 2026 (UTC)

Nobody gave a firm warning that would force a response. Since I have seen no LLM evidence in that article I did not, and wont. If you feel there was please go ahead with a level 2 or 3. Ldm1954 (talk) 00:52, 29 April 2026 (UTC)

The fact that nobody had given a firm warning that would force a response is another indication that this was too early to decide which of the scenarios in your analogy we are in. Stepwise Continuous Dysfunction (talk) 01:09, 29 April 2026 (UTC)

The person who is of the opinion that there is large scale LLM is Gnomingstuff. It makes sense that they should have given a warning. Since I saw no such evidence I did not until there was evidence at a completely different AfD which triggered my warning -- but only for that AfD.

I had already done the NPP review and minor editing without evidence of LLM, the LLM tag was added after I had marked it as reviewed. Ldm1954 (talk) 01:27, 29 April 2026 (UTC)

But you're completely fine with giving people warnings for tagging things for speedy deletion when they didn't. The irony of going through this much hand-wringing about someone possibly putting an AI-generated warning on an article with unclear grounds, when we have an actual, concrete, non-ambiguous example of a false warning here, is staggering. Gnomingstuff (talk) 01:29, 29 April 2026 (UTC)

...so let me get this straight.

Someone completely independent of all this flagged them for using AI in January. Several months later, someone else (me) flagged them for using AI. Other people experienced in AI cleanup also agreed that the text seems likely to be AI-generated. And now, the editor involved has now admitted they are indeed using AI for at least some things.

At what point, exactly, does WP:DUCK come into play? Gnomingstuff (talk) 01:26, 29 April 2026 (UTC)

@Gnomingstuff, you're welcome to use whatever heuristic you like to tag articles you believe are AI-generated. But when someone is done checking the article, they are then free to remove the tag. Maintenance tags are intended to draw editor attention to a problem so the problem can be checked and/or solved. When that's done, they're supposed to be removed. -- asilvering (talk) 13:18, 29 April 2026 (UTC)

As I said earlier, please don't ping me. Gnomingstuff (talk) 14:18, 29 April 2026 (UTC)

Returning to the topic (maybe), to give a related example I recently dePROD'd a page where the rationale provided was No sigcov, likely AI, possibly UPE. To me this falls under WP:ASPERSIONS and should not be done without decent evidence, innocent until proven guilty. I think there is a tricky issue here, but maybe I am alone in this view. Ldm1954 (talk) 06:58, 28 April 2026 (UTC)

@Ldm1954, without a link to the article in question, it's hard to say whether that's been done without decent evidence. -- asilvering (talk) 16:16, 28 April 2026 (UTC)

I don't think that falls under "aspersions", any more than an edit summary like "rv spam". Stepwise Continuous Dysfunction (talk) 22:56, 28 April 2026 (UTC)

In Wikipedia_talk:New_pages_patrol/Reviewers/Archive_53 is a discussion "Dealing with LLM tags" that I started on a similar theme. The main point there was it would help future editors if the LLM tag had an explanation with it, perhaps even just in the edit summary. I still agree with that position. The main counterargument was that LLM use is so obvious the tag doesn’t need an explanation, but that’s not been my experience. On a different point in the above discussion, are we saying that any editor caught abusing LLM is effectively implicitly banned from creating articles? If so, that should be made explicit somewhere, and normal banning procedures followed. --Northernhenge (talk) 18:51, 28 April 2026 (UTC)

I don't think anybody has said that. However, new articles by a user who has been caught using LLM and not responded to messages regarding this, should be treated as highly suspect. I'm not personally aware of the details of the current case, but if somebody has definitely used LLM the main concern is to stop them polluting the water with more shite, rather than any theoretical assumption of innocence. So, non-responsive AI-article creators should get everything they have created preventively draftified or stubbified until they respond.Boynamedsue (talk) 19:18, 28 April 2026 (UTC)

I agree. Looking at Wikipedia:Banning policy makes me think we need something like a topic ban, but banning the nonresponsive LLM-user from creating articles. The policy includes circumstances where the article should be deleted. --Northernhenge (talk) 08:42, 29 April 2026 (UTC)

@Northernhenge, we don't need any kind of topic ban for this. We simply block. -- asilvering (talk) 13:13, 29 April 2026 (UTC)

Part of the discussion is about the situation where someone used LLM in the past to create an article – does that mean we should assume their subsequent articles are also likely to be LLM-generated and always tag them? That discussion seems to assume the LLM user wasn’t blocked, but remains able to create articles. I’m concerned that NPP reviewers might be encouraged to tag all of an editor’s creations as LLM, without any process underpinning that and without any consistent means of notifying the errant editor that they must respond to challenges before they create any more articles. Bringing all this within an existing procedure, such as the banning policy, makes the situation clearer to everyone involved. --Northernhenge (talk) 13:29, 29 April 2026 (UTC)

Great news! There's an RfC about that literally right now!

I guess I also don't understand why people are so much more horrified about the prospect of an article tagged for potentially containing AI-generated content, than about an article actually containing it. Gnomingstuff (talk) 14:36, 29 April 2026 (UTC)

@Northernhenge, if another editor is unresponsive, it doesn't matter whether they're using LLMs or not - we routinely hand out blocks for failure to communicate. It seems to me a perfectly reasonable assumption that an editor who has used LLMs to generate articles, and is not responding to queries, is still using LLMs. I'd hope that any tagging editor would at least be giving the article a skim before applying the tag. -- asilvering (talk) 17:08, 29 April 2026 (UTC)

Since December/January or so I have left an explanation on all of my AI tagging, including the one in question here. Once again, you are free to verify this for yourself. Gnomingstuff (talk) 21:10, 28 April 2026 (UTC)

New pages patrol May 2026 Backlog drive

May 2026 Backlog Drive | New pages patrol
On 1 May 2026, a one-month backlog drive for New Page Patrol will begin. Barnstars will be awarded based on the number of articles patrolled. Barnstars will also be granted for re-reviewing articles previously reviewed by other patrollers during the drive. Each review will earn 1 point. Interested in taking part? Sign up here.
You're receiving this message because you are a new page patroller. To opt-out of future mailings, please remove yourself here.

MediaWiki message delivery (talk) 11:25, 29 April 2026 (UTC)

Redirect Backlog

My goal with this discussion is to get a general impression on how others feel, not already change anything. I was recently thinking about the length of our backlog, specifically the giant redirect backlog and more sustainable solutions. The Wikipedia:RWHITELIST caught my attention. My main two questions were, how many redirects in the queue were created by New Page Patrollers and how often do articles created by NPP end up at RFD? I'd think that given the experience we expect of NPPs and the rather simplistic nature of redirecs that we could be trusted with creating redirects without additional scrutiny. Important I'd not be proposing to autopatrol all of us, only redirect creations could be affected. Happy to hear your thoughts ~ Squawk7700 (talk) 23:48, 29 April 2026 (UTC)

I had a look at WP:RFD#DELETE to see if any of the problems were likely to be caused by mistake / honest difference of opinion, and the only obvious one (to me) was number 10, "If the redirect could plausibly be expanded into an article, and the target article contains virtually no information on the subject." The other reasons were either broadly technical, or else cleaning up possible mischief. Assuming NPP reviewers are not disruptive, even on bad days, I therefore don't have a problem with the suggestion. --Northernhenge (talk) 08:14, 30 April 2026 (UTC)

Yep, this sounds like a decent idea. Toadspike [Talk] 08:48, 30 April 2026 (UTC)

I'd say only Rosguill has more experience reviewing redirects and managing WP:RWHITELIST than I do, but speaking based on my experience, I'm not a fan of this idea, and I don't think it would have much of an impact. NPPers are often focused on articles, and many of them aren't page movers, but newer NPPers and those who are rusty have, in the past, created redirects from moves that I've needed to revert enough times to make me dislike this idea. I've also had to talk to NPPers over the years about what constitutes a proper redirect, and I don't think the user group is a suitable exception.

With that said, it could make sense to run a quarry query for NPPers who are not autopatrolled but have created over 100 redirects, then run that against the list of RAL users to see if there are some suggestions that can be made. Hey man im josh (talk) 12:18, 30 April 2026 (UTC)

Thanks for your input, those experiences are definitely worth considering, especially since I haven't had the chance to make them since I am rather new. With you touching the subjects of PM's, would you be open to RWHITELIST them? Or are there not considered caveats as well? Squawk7700 (talk) 16:51, 30 April 2026 (UTC)

Redirects created as the result of page moves, by page movers, are already automatically patrolled @Squawk7700, while page movers themselves are not. I've seen some PMs create some bad redirects, and they may be great about moving pages upon request, or fixing some small issues, but sometimes creating whole new redirects, they may not be the best at. Hey man im josh (talk) 17:01, 30 April 2026 (UTC)

Huh, I didn't know... How dumb of me, well that sorts that as well, disregard the previous comment. Squawk7700 (talk) 17:18, 30 April 2026 (UTC)

It does seem as if the redirect autopatrol list is pretty underpopulated in general. The last application was 3+ months ago, while WP:PERM/AP itself gets multiple applications every week. ScalarFactor (talk) 16:54, 30 April 2026 (UTC)

I think you'd be surprised how many people make a lot of flawless redirects that aren't already on the RAL or autopatrolled. Sometimes there's so few redirect creations, such as one or two a month over the course of 5+ years, that we don't even notice people reaching the level to qualify for RAL. Hey man im josh (talk) 17:02, 30 April 2026 (UTC)

Would it make sense to try "advertising" it a bit more? The header on WP:PERM/AP mentions that redirects (and dabpages) aren't counted for the 25-article-minimum for autopatrolled, but saying "if you create a lot of good redirects, consider applying at WP:RWHITELIST instead" could be helpful. (I'm aware that it is mentioned under the "handled elswhere" header but frankly I'm pretty sure a lot of people don't look at that, they just scroll down to the permissions.) ScalarFactor (talk) 17:07, 30 April 2026 (UTC)

Typically most creators don't necessarily care whether their content is marked as patrolled or not, and if they do, it's usually people wanting their articles to appear on Google before 90 days (since the creation of the article). Really what we'd want is NPPers making the recommendations because they'll be the ones who care about getting the redirect backlog down. MPGuy2824 is the most significant redirect reviewer since I became less active and they're already aware of this page, and they've made a number of recommendations in the past. I think it's just a matter of not having a ton of redirect reviewers, and if we do get redirect reviewers, making sure they are aware of the RAL. Hey man im josh (talk) 17:12, 30 April 2026 (UTC)

I don't panic about RFD having a backlog because 1) it's not nearly as important as the article backlog (the article backlog gets hit by UPEs who, if left unchecked, would corrupt the encyclopedia with their lack of NPOV), and 2) redirects not patrolled in 6 months get autopatrolled by the software, which keeps the backlog from getting too high. –Novem Linguae (talk) 13:27, 1 May 2026 (UTC)

My perspective on the topic: There are multiple types of revisions needed for the backlog redirects, after reviewing about a dozen of them:

• MOST lack an R-template, which is kind of important in my mind, but which many other think is unnecessary
• Some point at articles where they should point at extant sections.
• Some point at articles where they should (for convenience) point at anchors ... and I've been adding those anchors to the target articles in a couple of cases

I think there is a place for redirect-NPP, but it might be useful to split redirects from articles, a separate NPP list and a different set of suggestions for redirects emerging from NPP as 'reviewed'.

Regards --User:Ceyockey (talk to me) 03:07, 9 May 2026 (UTC)

WP:NPPSCHOOL is looking for trainers

Hello friends. Looks like we only have one WP:NPPSCHOOL training spot available (thanks CFA!) So it might be a good idea to recruit more volunteers to teach NPP school.

The ideal trainer is someone that has completed NPPSCHOOL as a student. However anyone who is comfortable with notability (WP:GNG, WP:SNG) can become a trainer.

The way it works is you add yourself to the mentor list, then interested students contact you on your user talk asking to take one of your slots. You agree, edit the mentor list to adjust your # of slots, create a user subpage using the standard curriciulum, the mentee fills out a section, you grade it. After that round of back and forth, you add the next section, the fill it out, you grade it. etc. etc. It takes around 1-3 months, but is done casually, with you and the mentee posting every couple of days.

Anyway, if you want to step up and help out, please 1) add yourself to the mentor list, then 2) DM me or email me and I'll send you the curriculum. (I'm not posting the curriculum here because there's been some cheating in the past involving students copying other student's answers, so would like to keep it low visibility.)

Thanks!

cc @Barkeep49, Asilvering, and CoconutOctopus: –Novem Linguae (talk) 22:08, 1 May 2026 (UTC)

Would I qualify as trainer? Vanderwaalforces (talk) 22:27, 1 May 2026 (UTC)

Yes. -- asilvering (talk) 10:41, 2 May 2026 (UTC)

Sure. Go ahead and add yourself to the list, and I'll DM you the curriculum. –Novem Linguae (talk) 14:05, 2 May 2026 (UTC)

NPP School was among my most rewarding activiites as an active NPP'er. For those who are sufficiently skilled with NPP and interested in mentoring I highly recommend it. Best, Barkeep49 (talk) 22:46, 1 May 2026 (UTC)

Year x in country y

What are the notability criteria for articles with titles like "1977 in San Marino"? Are they like list articles where the user has little obligation to prove the notability of the category, or do some other criteria apply?Boynamedsue (talk) 05:35, 2 May 2026 (UTC)

That's a good question. I think they'd be lists, and thus fall under WP:NLIST, but I'm not certain. Toadspike [Talk] 22:25, 2 May 2026 (UTC)

Assuming it does fall under NLIST, it's also not immediately clear to me how one would go about analyzing whether a list of this nature meets that criteria. I'm assuming most countries have, at some point, had a chronological histories written of them, would that be sufficient to fulfill this criteria? An argument could also be made that such lists are valid as a navigational aid. Might be worth starting a discussion at WT:N to get more opinions on this? 🌸⁠wasianpower⁠🌸 (talk • contribs) 01:19, 3 May 2026 (UTC)

Yeah, it's quite a difficult one, especially with older years where the category isn't easy to establish with sources.Boynamedsue (talk) 09:44, 3 May 2026 (UTC)

Maybe look at some recent AFDs from this area, see who participates in them, and ping whoever looks knowledgeable for advice on what the de facto customs/criteria are for these. –Novem Linguae (talk) 10:11, 3 May 2026 (UTC)

How do you log your deletions?

I recently got NPR rights, prior to that I was only using Twinkle for deletions and it automatically logged everything to my userspace. Now I can tag deletions using the page curation toolbar. Of course, that automatically counts for the backlog drive, but I'm not sure it's a better experience. So, I was wondering what other reviewers use. For deletion tagging, do you use Twinkle, the toolbar, or something else? StartOkayStop (talk) 06:56, 2 May 2026 (UTC)

In most cases deletion nominations with the toolbar would be recorded together in the Page curation log. Vestrian24Bio 10:49, 2 May 2026 (UTC)

I use Twinkle because the user experience is superior. Dclemens1971 (talk) 16:59, 2 May 2026 (UTC)

Definitely Twinkle. Easier to keep track of when you have it generate logs, in case you need to look back at any of them. • Quinn (talk • it/its) 01:35, 3 May 2026 (UTC)

What are the notability criteria for laws and bills and so on?

King's Revenues Act 1515 is very short, it really just say this exists. But I realise I have no idea if there are specific notability criteria for laws? And of course there might be other sources. Lijil (talk) 14:37, 4 May 2026 (UTC)

I wouldn't be so certain that a 16th-century act of parliament has much in the way of sourcing, and I'm unaware of any SNG for acts of parliament. -- asilvering (talk) 15:00, 4 May 2026 (UTC)

Just WP:GNG I think. Although the sources simply need to exist and don't need to be in the article (WP:BEFORE). –Novem Linguae (talk) 20:02, 4 May 2026 (UTC)

Yeah. I'm seeing sources about "King's Revenues" but not about a specific act in 1515. Might look a bit more tomorrow otherwise I guess AfD. I'm starting to feel so mean about all the AfDs though. Lijil (talk) 22:14, 4 May 2026 (UTC)

If it was clearly part of a series of articles, I’d suggest draftifying, but the act that repealed it was a general-purpose mass-tidying so not a successor in any meaningful sense. AfD is a good place to debate the actual article, so not necessarily something to feel mean about. Maybe somebody with subject-matter expertise could create King's Revenues and redirect 1515 to there, but this isn’t the place for that discussion. --Northernhenge (talk) 09:03, 5 May 2026 (UTC)

Fair enough, thanks for the feedback! Lijil (talk) 14:33, 5 May 2026 (UTC)

User talk notification edit summary

Tracked in Phabricator
Task T414521

This seems missing - see this diff. Ping Aafi, when responding. ─ Aafī on Mobile ^(talk) 18:59, 11 May 2026 (UTC)

Possible Orphan Label

I was recently granted new pages patroller. There have been some articles tagged in the New Pages Feed as possibly orphaned that I have hyperlinked to other articles. However, the "possible orphan" label on the New Pages Feed remains on those articles days later (for example, Aviteur). Would it be possible for the "possible orphan" label on the New Pages Feed to be removed after an article is hyperlinked to other articles? EaglesFan37 (talk) 15:37, 12 May 2026 (UTC)

An "orphan" is a page with no wikilinks to it. Did you link other articles to that one -- your message implies the other way around. Ldm1954 (talk) 16:11, 12 May 2026 (UTC)

@Ldm1954 I did (sorry if that was unclear. I linked Aviteur to Patricia Gucci on May 6. On the New Pages Feed, it still has a "Possible Orphan", Aviteur still has a "Possible Orphan" label. EaglesFan37 (talk) 16:17, 12 May 2026 (UTC)

In your question, I'll assume that you meant "that I have hyperlinked from other articles". You need to do a null edit (or any edit) on the article in the NPP queue. That should update all the automatically generated issues for the page. -MPGuy2824 (talk) 16:16, 12 May 2026 (UTC)

@MPGuy2824 I did mean from. Sorry for any confusion. The null edit worked. Thank you! EaglesFan37 (talk) 16:21, 12 May 2026 (UTC)

New Page Patrol Newsletter - May 2026

Hello New pages patrol,

New Page Review queue November 2025 - May 2026

Backlog update

At the time of this message, there are 15,282 articles and 32,951 redirects awaiting review.

After the January–February drive the article backlog was reduced to 15,179 articles and the redirect backlog to 19,053 respectively. Great job! However, both queues are growing rapidly and any additional reviews are highly appreciated.

2024 and 2025 NPP Awards

JTtheOG was selected as the NPP reviewer of the year for both 2024 and 2025, for reviewing the most articles amongst all reviewers.

Hey man im josh and MPGuy2824 won the Redirect Ninja Master Award for 2024 and 2025 respectively, for reviewing the most redirects.

Overall in 2024, one Platinum, two Gold, eight Silver, 12 Bronze and 45 Iron Barnstars were awarded. Additionally, 66 reviewers got the NPP barnstar for doing more than 100 reviews through the year. In 2025, one Platinum, ten Silver, 13 Bronze and 38 Iron Barnstars were awarded. Additionally, 38 reviewers got the NPP barnstar for doing more than 100 reviews through the year.

BoyTheKingCanDance, Rosiestep, SunDawn, and Vanderwaalforces were inducted into the NPP Hall of Fame for having two separate years of 2,000+ article reviews.

January–February backlog drive

The experimental two-month long backlog drive concluded with 183 reviewers patrolling over 27,761 articles and 35,309 redirects, earning over 36,836 points. Congratulations to JTtheOG, who achieved first place with 6,484.6 points in this drive.

May backlog drive

An article-only backlog drive is currently underway. We are hoping to make a big dent in the backlog. You can read more about it or join at Wikipedia:New pages patrol/Backlog drives/May 2026.

PageTriage

An attempt was made to get the New Pages Feed to sort by date marked as reviewed instead of date created. However we had to revert it due to bugs. We may try again in the future. You can subscribe to the Phabricator ticket if you're interested in following along.

Reminders:

• You can access live chat with patrollers on the New Page Patrol Discord.
• Consider adding the project discussion page to your watchlist.
• To opt out of future mailings, please remove yourself here.

MediaWiki message delivery (talk) 23:38, 14 May 2026 (UTC)

This says "both queues are growing rapidly", but the image shows the article queue decreasing (which is of course impressive but still contradicts the message). ~ Jenson (SilverLocust 💬) 23:48, 14 May 2026 (UTC)

Ouch - I'll have to apologize for this one, I originally wrote that part before the backlog drive and then revised it again, it made sense to me when I wrote it but obviously it doesn't really. Best, Squawk7700 (talk) 20:55, 16 May 2026 (UTC)

Species Helper errors

I'm getting this error about 50% of the time when using Species Helper for the last couple of weeks - just wondering does it affect anyone else? Just pinging Novem Linguae in case there is anything amiss in the script. Thank you Josey Wales ^Parley 09:45, 15 May 2026 (UTC)

@EatingCarBatteries also reported this at User talk:Novem Linguae#Garbled response using SpeciesHelper. Something probably broke somewhere. I'll need to carve out some time to debug it. –Novem Linguae (talk) 06:08, 16 May 2026 (UTC)

Failing page in screenshot: Macratria durrelli. –Novem Linguae (talk) 06:29, 16 May 2026 (UTC)

Should be Fixed now. Thanks for reporting. –Novem Linguae (talk) 06:30, 16 May 2026 (UTC)

@User:Novem LinguaeThanks for all your work on this Josey Wales ^Parley 08:55, 16 May 2026 (UTC)

LDS sources and notability for an LDS temple

What do we think about this article? Cagayan de Oro Philippines Temple

There is decent sourcing but it is all LDS newspapers or websites. Are we having it or not?--Boynamedsue (talk) 18:53, 15 May 2026 (UTC)

Excessively detailed pages

I've come across a user named @Muffinbutt1985 who has been making excessively detailed Wikipedia pages that don't follow WP:TOOMUCH. Yusra Hussien and Samya Dirie, Portsmouth terrorist cell, Yusra Ismail, Mannan family terrorist cell, and Zafirr Golamaully. CostalCal (talk) 02:24, 18 May 2026 (UTC)

What can I do to make them Sound more neutral? I’m autistic with a special interest in it Muffinbutt1985 (talk) 02:25, 18 May 2026 (UTC)

@Muffinbutt1985 - Although I do believe that it mostly retains a neutral point of view. I've changed the tag. And I do thank you for creating these articles. However, it's still obvious that these articles go into way too much detail, per WP:TOOMUCH. Wikipedia is only supposed to be a summary. CostalCal (talk) 03:21, 18 May 2026 (UTC)

I will work on removing the excess details. Muffinbutt1985 (talk) 03:22, 18 May 2026 (UTC)

Thanks! CostalCal (talk) 03:24, 18 May 2026 (UTC)

I’ve worked on removing excessive detail from a couple of the entries, I removed the “excessive detail” box from the ones I believe I have fixed. Will be working on some others later today. Muffinbutt1985 (talk) 22:30, 18 May 2026 (UTC)

Unable to move article due to page lock

I reviewed the AFC article Draft:Goyslop and was going to accept it through AFC. I'm unable accept it through AFC since a soft redirect for the page is globally locked. How should I go about accepting this page? EaglesFan37 (talk) 23:27, 18 May 2026 (UTC)

I believe this would count as a request to decrease page protection, which would make the protecting administrator the first person to contact. ScalarFactor (talk) 23:41, 18 May 2026 (UTC)

@ScalarFactor It looks like @Asilvering already G6ed the soft redirect, but I will do as you suggested in the future! EaglesFan37 (talk) 23:52, 18 May 2026 (UTC)

@EaglesFan37, I've G6'd, and you're clear to accep the draft. -- asilvering (talk) 23:47, 18 May 2026 (UTC)

@Asilvering Thank you! EaglesFan37 (talk) 23:51, 18 May 2026 (UTC)

A small terminology correction. I don't think there's a protection level called "globally locked". It might be more accurate to say "protected", since "globally locked" is a term that usually refers to user accounts (WP:GLOCK). Hope that helps! –Novem Linguae (talk) 00:22, 19 May 2026 (UTC)

@Novem Linguae I meant fully locked, not sure why I wrote globally locked lol. Thank you! EaglesFan37 (talk) 00:44, 19 May 2026 (UTC)