Contributors frequently propose new (or expansions of existing) criteria for speedy deletion. Please bear in mind that CSD criteria require careful wording, and in particular, need to be
Objective:Most reasonable people should be able toagree whether a page meets the criterion. Often this requires making the criterion very specific.
Uncontestable:It must be the case that almostall pages thatcould be deleted using the rule,should be deleted, according to consensus. CSD should cover only situations where there is a strong precedent for deletion. Remember that a rule may be used in a way you don't expect, unless you word it carefully.
Frequent:Speedy deletion is intended primarily as a means of reducing load on other deletion methods such asWikipedia:Articles for deletion andWikipedia:Proposed deletion. These processes are more discriminating because they treat articles case-by-case, and involve many points of view; CSD sacrifices these advantages in favor of speed and efficiency. If a situation arises only rarely, it's probably easier, simpler, and fairer to delete it with one of the other methods instead. This also keeps CSD as simple and easy to remember as possible, and avoidsinstruction creep.
Nonredundant:If the deletion can be accomplished using a reasonable interpretation of an existing rule, just use that. If this application of that rule is contested, consider discussing and/or clarifying it. New rules should be proposed only to cover situations that cannot be speedily deleted otherwise.
If you do have a proposal that you believe passes these guidelines, please feel free to propose it on this discussion page. Be prepared to offer evidence of these points and to refine your criterion if necessary. Consider explaining how it meets these criteria when you propose it. Do not, on the other hand, add it unilaterally to the CSD page.
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.
Emojis: This may include usage ofEmojis with no encyclopedic value, only if used in the mainspace and draftspace page at the beginning of paragraphs or in place of bullet points.
Support as proposer. As also noted atWikipedia:Signs of AI writing#Emoji, LLMs are notorious for leaving emojis in their output. Such articles can be safely deleted because human editors will most probably not include emojis in their writing, unlike the possibility with markdown (older RfC).~/Bunnypranav:<ping>12:22, 27 August 2025 (UTC)[reply]
Support, with caveat. I agree that this is highly indicative of AI and that it is very unlikely to be used by a normal human editor. I acknowledge that it ispossible that a human editor reviewing LLM output would leave emojis in articles, but find it unlikely. There is a bit of a complication here in that this is a G criterion and not an A one. I wouldn't approve of someone G15'ing a userpage for this reason, for example - someone who used AI to create a quick userpage shouldn't have it deleted as "unreviewed by a human" if they used emoji bullet points to describe their editing interests. In drafts or articles, however, I agree that this demonstrates lack of human review and is thus in line with the spirit of G15. --asilvering (talk)18:48, 27 August 2025 (UTC)[reply]
I have seen GreenLipstickLesbian's analysis below, but "used in the mainspace and draftspace page at the beginning of paragraphs or in place of bullet points" is sufficiently indicative and discriminate. —Alalch E.14:08, 2 September 2025 (UTC)[reply]
Oppose per my comment and example below; not an unambiguous LLM tell. In fact, I'm pretty sure humans are more likely to use emojis than markdown, and that proposal is getting pretty resoundingly vetoed. (Just to talk about biases, I think the fact that Wikipedia editors tend to be very techy makes us think markdown is more common than it is. It's certainly not more common than emojis, anyway.)GreenLipstickLesbian💌🦋21:50, 27 August 2025 (UTC)[reply]
Oppose At least not without further workshopping. Examples given below have demonstrated that LLMs are not unique in their use of emojis. I think OutsideNormality provides a good starting point for narrowing down the proposed criterion. I am also worried that this indicator will soon be obsolete. According to OpenAI, GPT-5 now uses "fewer unnecessary emojis", and I expect that other LLM makers will follow suit.Catalk to me!07:24, 28 August 2025 (UTC)[reply]
Conditional Support Per GreenLipstickLesbian, emojis are even more likely to be used by novice human editors than markdown, the latter of which appears en route to rejection as a G15 criterion. Even while knowing that they are unprofessional for article text, I use them in plenty of edit summaries as tone indicators. Accordingly, I only support this criterion if limited to use at the beginning of paragraphs or in place of bullet points.ViridianPenguin🐧 (💬)11:30, 28 August 2025 (UTC)[reply]
Note that the proposed criteria only applies if emojis are used "in a mainspace or draftspace page at the beginning of paragraphs or in place of bullet points." No offense or rebuke intended to ViridianPenguin. I would insert a smiley with cold sweat running off brow emoji here if I remembered how to do so.--FeralOink (talk) 08:01, 9 September 2025 (UTC)--FeralOink (talk)08:01, 9 September 2025 (UTC)[reply]
Support per various comments above - support conditional on it only applying to mainspace and draftspace, and only when the emojis are used as bullet points / at paragraph starts.CoconutOctopustalk11:38, 28 August 2025 (UTC)[reply]
Oppose - changing my !vote after seeing GreenLipstickLesbian's rundown.Conditional oppose: Emojis are more indicative of being on a mobile phone than anything, and "no encyclopedic value" is more subjective than I would prefer for a speedy criterion. I would however support narrower criteria, like emoji bullet points.Gnomingstuff (talk)16:18, 28 August 2025 (UTC)[reply]
Oppose - in my experiments with CHAT GPT I have yet to see it use emojis. In my experience with humans, I have seen them use emojis. I also have had some browsers convert typos to emojis automatically. I don't think this is an obvious AI-tell that raises to the level of being added to the CSD. ~ONUnicorn(Talk|Contribs)problem solving16:22, 28 August 2025 (UTC)[reply]
Oppose. Firstly this is not a reliable indicator of unreviewed LLM output; secondly no evidence has been presented that this is needed: proponents should provide examples of pages that should be deleted, but cannot be deleted under the existing speedy deletion criteria (not just G15), but could be if something related to emojis were adoptedand some evidence that these occur frequently.Thryduulf (talk)22:16, 28 August 2025 (UTC)[reply]
Oppose - literalyears of debates over content with emoji show that people (particularly younger generations) have been using emoji in their natural writing formuch longer than LLM content generators have been a thing, and many inexperienced editors write in the way that they naturally communicate rather than the fairly polished academic style of Wikipedia. Also per Thryduulf: this fails the "frequent" bullet of the CSD expansion criteria as no example use case has been identified at all.Ivanvector (Talk/Edits)22:30, 28 August 2025 (UTC)[reply]
Weak support, providedat the beginning of paragraphs or in place of bullet points. Emojis sprinkled willy-nilly through the text aren't an indicator of AI (though they will reflect other things about the editor), but the bullet point things is a hallmark. I would urge caution if this is theonly criterion met though, and the text doesn't otherwise show signs of AI – the writer may just be an imbecile rather than an LLM.Cremastra (talk·contribs)22:35, 28 August 2025 (UTC)[reply]
I wouldoppose for failing the frequency criterion. I would also suggest freezing G15 for a few months with no further expansions, while it beds in.Stifle (talk)08:06, 4 September 2025 (UTC)[reply]
Oppose, not a sure tell for being AI, and not a good way to treat human volunteers. Emojis are easily removed through normal editing.ϢereSpielChequers18:46, 6 September 2025 (UTC)[reply]
Oppose - I don't think it's a clear enough tell of LLM misuse, doesn't seem frequent enough to be an especial problem that can't be dealt with through other means. Cheers,SunloungerFrog (talk)06:20, 7 September 2025 (UTC)[reply]
Support - Although I do SUPPORT, a citation that gives evidence of LLM output being prone to use emojis would be helpful. I perused the linkedWP:AISIGNS which supports this proposal. It also states in the second-to-last paragraph of the lead that "The speedy deletion policy criterion G15 (LLM-generated pages without human review) is limited to the most objective and least contestable indications that the page's content was generated by an LLM. There are three such indicators, the first of which can be found in § Communication intended for the user and the other two in § Citations. The other signs, though they may indeed indicate AI use, are not sufficient for speedy deletion." It is anWP:ADVICEPAGE only, not WP policy.--FeralOink (talk)07:52, 9 September 2025 (UTC)[reply]
Support Any human who has read a few Wikipedia articles knows they never use emoji (unless its a page about an emoji or in a quote) so if an AI makes a wikipedia article without being told its going on wikipedia, they wont follow WP:TONEmicroTato(talk)(contribs)— Precedingundated comment added17:24, 9 September 2025 (UTC)[reply]
I wonder if we all have the same idea about what an emoji is. How many of these bullets are emojis?
At a glance on my device (Firefox on xubuntu linux) I wouldn't identify any of them as emoji but as normal text bullets. I could accept an argument thatmaybe D is an emoji, but what is the dividing line between an non-alphabetic (or equivalent in other scripts), non-punctuation character and an emoji? If we go by the unicode definitions, then I believe that none of the above are emoji. Indeed the only emoji used in this discussion are (in no particular order): 🐧💌🦋🐄🫘💕⭕✅🔹. The first 5 of those appear only in signatures, the 6th appears to be random text decoration in a human-written comment and the 7th is given as an example of legitimate use on another wiki.Thryduulf (talk)21:54, 9 September 2025 (UTC)[reply]
I just used text selection to determine if the examples are emoji or not. If the bullet turned white when I selected it, it was being rendered in plain text. If it stays black (or whatever colour it originally was) then it was being rendered as an emoji. That's also why I specified my operating system and browser, because emojis render differently on different devices.OutsideNormality (talk)01:09, 17 September 2025 (UTC)[reply]
None of them change color on my laptop (MacOS). I doubt that anyone here thinks that Item C (square dot emoji) is actually more obnoxious in an article than Item H (heart-shaped dingbat)?
IMO since we don't have a shared understanding of which characters are emojis, and since we do have a l-o-n-g history of people typing things in word processing documents and then copy/pasting the results – something that's particularly noticeable when there's a bullet list that uses non-wikitext formatting – we really shouldn't be setting up a rule that says "I get to delete your contributionwithout discussion if you didn't use my preferred style of bullet".WhatamIdoing (talk)22:30, 24 September 2025 (UTC)[reply]
Oppose Yes, LLMs are notorious for including emojis in their output, but emoji usage alone isn’t a definitive sign of AI writing; new editors sometimes use emojis in their writing. Without other signs of AI writing, it’s not objective enough for speedy deletion.Rosaece ♡talk ♡contribs10:53, 17 September 2025 (UTC)[reply]
I don't think new editors use emojiat the beginning of paragraphs or in place of bullet points. Note that this proposal isn't so broad as to include "emoji usage alone".FaviFake (talk)14:34, 17 September 2025 (UTC)[reply]
Support as someone who always has and always will take aWP:TNT approach to this sort of thing. Remember that emojis alone aren't grounds for speedy deletion; this proposal will simply allow them as requisite evidence to call a spade a spade and kill a blatantly AI-generated page with fire under G15 (which isn't hard without any of these tells).Passengerpigeon (talk)21:22, 17 September 2025 (UTC)[reply]
Oppose. Faulty or bad use of emojis does not provide a definite conclusion that an article is written by LLM. Speaking from my own experience, chatGPT (and maybe other LLMs) are using less emojis these days.✠SunDawn ✠Contact me!16:44, 21 September 2025 (UTC)[reply]
Oppose. Early this year I did a bit of online work for an AI-training company, so for a few months I was spending hours a day working with AI writeups, and I almost never saw emoji. I was instructed to watch out for emoji and flag them if they appeared, so I'm confident that output from the models with which I was working (I wasn't told which ones they were) would be highly unlikely to include emoji — therefore, to me, the presence of emoji indicates a human author.Nyttend (talk)20:37, 26 September 2025 (UTC)[reply]
What about something like "emojis for no encyclopedic reason". We shouldn't prohibit the use of it in articles likeEmoji, where they serve a genuine purpose.Kovcszaln6 (talk)12:57, 27 August 2025 (UTC)[reply]
Searching for these in draftspace is somewhat difficult, thanks to our tolerance of emoji in signatures and the ill-advised practice of putting{{AfC comment}} on the pages themselves instead of talk, and separately the use of emoji in reference titles (especially from the likes of Instagram); but for the sake of argument, anextreme low-effort search turned upDraft:Prowl transformer,Draft:Amy Sparkes,Draft:PATRICE MUBIAYI, andDraft:Dr. Arth White before I stopped looking. (Two of those are already clear speedies even without this proposed expansion, and I'd bet all four would be deleted if tagged; but please don't go and tag them just yet.) —Cryptic16:54, 27 August 2025 (UTC)[reply]
Second thought: even though emojisshouldn't be used in human-written encyclopediac content, it's certainly plausible that they might be, and I don't want G15 to slide towards "speedy delete anything that violates the MOS". If other signs of AI writing are present then emoji use is evidence that no reasonable human review happened but emojis aren't unambiguous signs of AI by themselves, are they? --LWGtalk13:45, 27 August 2025 (UTC)[reply]
Hmm. I do not think emojis are plausible in human text, unless at least some of it is copied from AI. The criterion specifies[...] removed by anyreasonable human review anyways, so this shouldn't go that extreme ofdeletion due to violation of MOS.~/Bunnypranav:<ping>14:19, 27 August 2025 (UTC)[reply]
What about narrowing it down only to emojis in front of section headers or as bullet points (✅ and 🔹)? They are the main places where AI uses them, and based on my quick read of your comment, humans may put it in the content/end of sentences (correct me if I'm wrong).~/Bunnypranav:<ping>14:43, 27 August 2025 (UTC)[reply]
I mean, there's a reason we've had an edit filter since 2015[11] to prevent new editors from adding emojis to articles. Anecdotally, I've seen people use emojis in ref names, pre-chat GPT. I can't dig out the example I'm thinking of but I remember seeing somebody doing that in pre-2020 page revisions. (And then being very sad when I discovered that somebody had removed them from the modern version). I get the feeling behind this proposal, but it needs a lot more workshopping to be viable. Here's a few examples of emojis/unicode stuff being used in articles, anywhere from constructively to "plausibly good-faith".
Probably a decent number of the quarter million hits on Filter 680, actually.[12]. Many of them probably aren't going to be worth saving, but very few are going to be unambiguously LLM generated, and more than a few will be false positives. But yeah, this does prove that humans plausibly add emojis to articles.
[13] and[14] for a couple seemingly good-faith "no, we don't want you to do that edit but it's not exactly wrong either" style hits.
[15] Edit filter person talking about how you have to exempt Japanese-related articles from emoji/unicode filters like this.
Planets in astrology uses a lot of decorative emojis.[20] I think they might be astrology symbols, but I really don't know. To an admin not versed in astrology, these have no value, and a human certainly should not have left them in past review. To one versed in astrology, I'm assuming these are plausible, but they shouldn't have been added anyway.
We sometimes have issues with people trying to add these to short descriptions (probably the low character count makes them think it's social media, idk)[21][22].
Sports editors gonna sports editor[23] + further conversation about how newbies try to mimic images/symbols by adding emojis to articles[24]
[25] 2018 warning from an admin to a new editor about using emojis in articles. It looks like the emoji was in a ref name to, so for anybody saying that "obviously, an admin wouldn't speedy a draft because they copied the source title... there you go! (Waves a small trout at @Acroterion for historic mistakes, and a very genuine thank you for the data point)
[27] This newbie drafting a list to potentially add to an article on the talk page, using emojis in a decorative manner (all wordings we've come up with would kill this list, despite it being the most obviously good-faith one so far)
Similarly,[28] is a 2020 edit using the blue dot as a bullet-point esque marking. It's still in the article and very much predates chatGPT.
In terms of what other wikis use (and therefore, what good-faith newbies will often add to our articles either by translation or by making the mistake of assuming what's okay there is okay here):
On jaWiki, ⭕ sometimes gets used for article likethis
Also, I can't find any diffs, but here are some other ways I can see people using emojis in a non-LLM generated fashion:
Some people add their edit summaries to the actual article, especially in draft space. Some people might use emojis in those comments. If somebody ran their draft through something like Grammarly, it's probably going to have LLM vibes and the fact that somebody added emojis should not make it speediable.
Stuff like the skull and cross emoji or hazard emojis ☢ are used in botany and chemistry to denote that plants or chemicals are scary, so it's plausible a newbie might try to add those symbols to article text in what would appear to be a decorative way.
flaggggggsssss.
I'd imagine that AfC editors tend to mostly see emojis in LLM-generated spammy draft, but I'm not convinced that if anybody had instead of making these additions to an existing article, started out a new article and included emojis in such a manner, then their work should be insta-vanquished. "encyclopedic value" is nebulous; I'd argue that very few of the examples I listed actually have any value, and I'd happily remove most of the emojis if I wasn't using them as examples. But no, I don't agree with the premise that emojis cannot be plausibly inserted by a human.GreenLipstickLesbian💌🦋21:47, 27 August 2025 (UTC)[reply]
Actually, as far as those "this emoji seems decorative/valueless to some editors but isn't really", this edit by a newbie was stopped in the edit filter because they added an unicode emoji heart thing[30]. They tried to get somebody to add the emoji for them, but were turned down[31] and later templated by an admin for MOS:DECOR issues relating to the edit.[32] What neither of the more experienced editors realized, however, despite the newbie explaining themselves and the fact there were already 25 other hearts on the page and that fact it's explained in the key, is that for some esoteric reason Wikipedia editors use the heart emoji to denote, I believe, mares.[33]. So as far as "encyclopedic value" of emojis, again, not as objective as it might appear on first glance. (Also hi @DreamRimmer and @Peaceray! I'm talking about some random edits you two made a few months ago, so sorry but also thank you providing me with an example!)GreenLipstickLesbian💌🦋00:09, 28 August 2025 (UTC)[reply]
Oh gosh, no worries at all, and mostly just a giant thank yo for manning the edit filter thing that reduces the number of headaches the rest of us editors have to deal with!💕 And yeah, I was born to a former horse-girl and a wanna-be cowboy, and it took even me a minute of thinking and reading to figure out the heart emoji thing.GreenLipstickLesbian💌🦋13:32, 28 August 2025 (UTC)[reply]
Haha, you are definitely a genius, and I have often noticed your great skills in handling copyright-related matters. I also want to thank you for all the wonderful work you do. –DreamRimmer■13:40, 28 August 2025 (UTC)[reply]
Thank you GreenLipstickLesbian for your detailed comment with diffs and examples! Based on your, and some others' suggestions, I have narrowed the proposal down specifically to bullets and start of paras. Wondering if you have a new take on this, or any further amendments for more clarity and less false-postives.~/Bunnypranav:<ping>13:56, 28 August 2025 (UTC)[reply]
Sorry for the delay, wanted to think about my response a bit more! I definately prefer the much more narrow scope and I'm a lot less opposed to it, however, I still have concerns. For example, I've just quickly foundthis userpage, created in 2020, that uses bullet point emojis instead of actual bullet points. The userpage or it's history is going to vanish at some point, but that's because it was a verbatim copy of this WHO Facebook post[34]. So obviously not LLM-generated, and this is an example of not one, but likely a small team of humans looking at text with emoji bullet points and going "yeah, seems good" followed by a new editor doing the exact same thing. I admint this is a slightly weird case, as in this page could be validly CSD-ed, but I don't think anybody wants this CSD-ed under G15.I'm also a little worried that some new editors (especially in poorly-thought through editathons or classes) might use something like ChatGPT to generate an outline, then fill it in with (possibly quite poorly written) human text. I can't remember the exact pages, but I'm pretty sure I saw a few drafts like that yesterday which were part LLM-placeholder with emoji titles, but the actual writing could have only been written by a human. (Part of the reason I didn't save the titles is that both articles sucked in a uniquely organic way, so I couldn't use them as examples as I don't like public shaming an unsuspecting human's writing if I can help it) And, I mean, as far as using AIs to edit Wikipedia.... I mean, I wouldn't recommend doing that for obvious reasons but it's something an otherwise competent human might plausibly do, it's not actually the worse way to use an LMM to create articles, and I think speedying those could be bitey and create a headache at DRV. I mean, a decent portion will suck and be horrible, but will they be horrible enough that I don't mind them acidentally getting cleared out under G15? I don't know. @BunnypranavGreenLipstickLesbian💌🦋21:51, 28 August 2025 (UTC)[reply]
As with my comment in#RfC: Including Markdown in G15 above, it's not emojis per se (or markdown, or dead links) that should make a page speedy-deleteable; it's subjectively LLM-generated content which isalso confirmed by one of these blatant signs. —Cryptic15:11, 27 August 2025 (UTC)[reply]
Per various suggestions above, I have amended the proposal to only be applicable to main and draft page, that too only in start of paragraphs of in place of bullet points. Open to rephrasing in any way for clarity.~/Bunnypranav:<ping>12:09, 28 August 2025 (UTC)[reply]
@Gnomingstuff Would you be willing to suggest an alternative for "no encyclopediac value"? Also, as of now, the proposal limits to bullet points and para beginnings, which I assume has no encyclopediac value by default.~/Bunnypranav:<ping>16:49, 28 August 2025 (UTC)[reply]
deleted on other grounds [and also had] other AI tells so its not an example of a criterion like this being needed. To demonstrate that there needs to be many examples of pages which cannot currently be deleted under existing criteria (including but not limited to G15).Thryduulf (talk)12:18, 31 August 2025 (UTC)[reply]
Speedy deletion is the exception not the rule. Only if there are sufficiently many instances of the same sorts of pages getting nominated andalways being deleted is speedy deletion appropriate. You also claim that it both had multiple AI tells and no other AI tells, it cannot be both.Thryduulf (talk)14:42, 31 August 2025 (UTC)[reply]
G11 is based on a page'sstate (the state of being literal spam), while G15 is based on the page creation process (i.e. that it has been generated and slopped onto Wikipedia with no human review), which is harder to determine objectively: at best, we can make specific criteria (the G15 criteria) which indicate that the page in question likely had no human review, but we can't time-travel to when the creator created the page. We can't just redefine G15 to beall potentially AI-generated content either, as not all AI-generated content is bad enough that we need to skip a deletion discussion and speedily delete it (not to mention false positives).OutsideNormality (talk)17:03, 31 August 2025 (UTC)[reply]
It provides an external link to the company's online store, the product's homepage, marketing info or specifications page, product purchasing order page .... yes, that's killing most of our articles on businesses.[35][36][37] And I have yet another thing to add to my list of "fine if an older editor adds it, blockable offense if it's in somebody's first ten edits"[38] (the block was discussed at length on AN so I don't feel too bad about talking about it here)GreenLipstickLesbian💌🦋22:11, 6 September 2025 (UTC)[reply]
I've long said that Wikipedia's rules apply less to editors like me than to some other editors. This makes some sense: For example, I am the all-time top editor ofWikipedia:External links, its talk page, andthe associated noticeboard,[39][40][41] so why shouldn't you trust me on a questionable URL? The plain fact is that there probably isn't anyone else on the planet who knows these rules better than I do, and fewer than I'd like who equal my knowledge. But there is also something slimy about a system, created by me and people like me, that says "we" are permitted to add article content that a new editor would be reverted or even punished for.WhatamIdoing (talk)17:54, 9 September 2025 (UTC)[reply]
The "tells" in the article text did not fall under G15. The AfD nomination argued that it was AI-generated, not because of any of the G15 criteria, but because there were websites and programs that seemed to have been completely made up. In other words, the nomination invoked "other, more subjective signs of LLM writing that [...] should not, on their own, serve as the sole basis for applying" G15.Stepwise Continuous Dysfunction (talk)22:31, 31 August 2025 (UTC)[reply]
I wasn't saying that it should have been speedied because of the problems that the nominator pointed out. I'm saying that it's an example of a page that was identified as AI-generated, which was also full of emojis in exactly the way that has been proposed as an indicator of AI-generated text.Stepwise Continuous Dysfunction (talk)20:10, 1 September 2025 (UTC)[reply]
If we cast too wide a net, we will catch articles that just have odd formatting. I recently sawan edit summary from an excellent and experienced editor, where the edit was largely a positive change, but it noted "I'm going to assume that this list of four came from AI." However, that list formatting was present in the Spanish Wikipedia articleback in 2018, and reflected the source (Nores et al., 2015) cited there.Rjjiii (talk)01:34, 4 September 2025 (UTC)[reply]
I think the G15 should instead acknowledge that bad formatting is a possible sign of LLM use, but it should not be the only indication, depending on context and the editor's experience withwikitext. I am also thinking about how G15 could note the existence ofWikipedia:Signs of AI writing while clarifying that it is an advisory guideline. --Minoa(talk)09:38, 9 September 2025 (UTC)[reply]
Oppose: This criterion only applies to absolutely, entirely unambiguous cases of LLM use. There are plausible reasons why a good-faith human editor would use emoji.JJPMaster (she/they)12:21, 11 September 2025 (UTC)[reply]
G15 is new, it seems to be working reasonably well thus far, but it's still bedding in and there's been a plethora of proposals, many on CENT, to add or tweak the criterion. I propose a 2-month moratorium on proposals to make further tweaks to the criterion while this bedding-in takes place, so that it isn't constantly changing during that time. After that point we will have a clearer view on what needs to be added or changed.Stifle (talk)08:10, 4 September 2025 (UTC)[reply]
This is not a bad idea. It takes time to figure out whether/how a rule is working, and if you change it every few weeks, you'll never figure it out.WhatamIdoing (talk)21:15, 6 September 2025 (UTC)[reply]
I think that this is a good idea. G15 now allows editors to deal with egregious cases of LLM misuse quickly. Rather than agonising over less obvious cases or criteria, let's let it settle down. Nothing to stop people from noting down personally cases that might allow the G15 umbrella to be expanded, for future discussion. Cheers,SunloungerFrog (talk)06:16, 7 September 2025 (UTC)[reply]
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.
This would hopefully add clarity and reduce the number of proposals to merge them. There would be both a small but not non-existent benefit and a small but not non-existent amount of disruption from such a change, I'm not sure which outweighs the other but given Cryptic said they wouldn't be opposed, I figured it's worth seeing what others think of it independent of any other suggestions (there would be no changes to the text of either criterion, or anything else other than the heading).Thryduulf (talk)15:06, 1 September 2025 (UTC)[reply]
I'm not sure this would do it. Those proposing the merge of the two are aware that one covers text, and the other covers files. What they are likely not aware of is the reason they are separate, and the past discussion.Cryptic's suggestion of a footnote would work better. The footnote could refer to the perennial proposal link. --Whpq (talk)12:32, 2 September 2025 (UTC)[reply]
I think that renaming the headings could help people who are looking at the Table of Contents or links to specific sections, especially if they don't understand our codes (G for General, F for Files).WhatamIdoing (talk)21:18, 6 September 2025 (UTC)[reply]
The current name for both is just "Unambiguous copyright infringement", only one of the suggestions includes parenthesis and your comment doesn't address the other.Thryduulf (talk)14:14, 3 October 2025 (UTC)[reply]
"it doesnot apply to articles about albums (these may be covered byCSD A9), products, books, films, TV programs, software, or other creative works, nor to entirespecies of animals."
On the other hand it says it covers "web content" so given that academic journals mostly nowadays are published online do they count as "books" or "web content"?
It does not apply. If a journal has no indication of importance, conduct aWP:BEFORE search and then either expand the article if you find useful information, orWP:PROD orWP:AfD if you don't. Keep in mind that journals are one of those things that the communityde facto maintains a pretty low notability bar for if they're reputable, even if there technically isn't a separate SNG for them. --Tamzin[cetacean needed](they|xe|🤷)14:15, 9 September 2025 (UTC)[reply]
(edit conflict) I'm pretty sure the reason that there are carveouts for albums/books/etc. is that in practice, a non-notable creative work is going to be a valid redirect title to the author (andWP:A9 is specifically to address the scenario where no article for a recording artist exists), and thus if not notable should almost always beWP:BLARed rather than deleted. Separately, I think in practice you'd have a hard time applying the "no indication of importance" speedy deletion criterion to a journal, as my sense is that even for very obscure journals, the claim of being a peer-reviewed is a plausible claim to importance (it may not be borne out, but that's a question for PROD/AfD, not CSD).signed,Rosguilltalk14:16, 9 September 2025 (UTC)[reply]
We could consider it and see if there's a consensus, but I'm mildly inclined to believe that it isn't really a problem that needs solving. As I laid out in my first comment, a page for a non-notable book is very likely to have a valid alternative to deletion. The same isn't quite true of magazines or journals thatcould be backed by an identifiable, notable institution, but could equally easily have a collective authorship that doesn't easily resolve to a single notable target (or alternatively, they could be published by an institution so noteworthy and famous and with so many publications that the one publication isn'tWP:DUE for mention at the institution article. Contra Tamzin (a little), I think we could imagine an edge case for a self-described journal that does not claim to be peer-reviewed and does not present any claims of impact or importance, which could meet A7 criteria--but such cases are extremely raresigned,Rosguilltalk14:27, 9 September 2025 (UTC)[reply]
A7 only applies to real people, individual animals, commercial or non-commercial organizations minus educational institutions, web content, and organized events, and as academic journals are neither of those things, A7 does not apply to academic journals. The list of what A7 is not is non-exhaustive and is entirely optional; it doesn't affect the policy norm at all and is just built-in commentary. —Alalch E.15:15, 9 September 2025 (UTC)[reply]
@Rosguill, How does the fact of being peer-reviewed (which is a norm for academic journals) count towards a journal's notability? It's not mentioned onWikipedia:NJOURNALS. Furthermore how is such a claim even verified? Any journal could claim they are peer-reviewed, but how do we establish the fact that it is (except for well-established journals)?Xpander (talk)05:52, 10 September 2025 (UTC)[reply]
Remember that a claim of significance is explicitly a lower standard than notability, and doesn't have to be correct, let alone verified as correct, just plausible. Being peer reviewed is a claim of significance.Thryduulf (talk)11:01, 10 September 2025 (UTC)[reply]
Xpander, you can establish this by asking someone withUlrichsweb access. We librarians use this massive database when making professional decisions about individual serials (it has an extremely high reputation for accuracy), and because it's relevant to scholars publishing in specific journals and scholars researching the significance of individual serials (whether journals or not), far more people than librarians would have access to it. I suspect most anglophone universities would have it, andWP:TWL orWP:RX might be able to help, too.Nyttend (talk)20:47, 26 September 2025 (UTC)[reply]
Why is "G6: Housekeeping and routine (non-controversial) cleanup" still in the delete dropdown?
I was curious as to when G6 changed from "Housekeeping" to "Technical deletions", particularly given how much some people complain about G6 being misused to the point whereWikipedia:What G6 is not is "See also"-ed at the top of the criterion and yet "Housekeeping and routine (non-controversial) cleanup" is so attractive for such misuse if someone hasn't carefully readWP:CSD lately. I was surprised to find out that change happened way back inJune 2008. If you really believe what that essay says, why hasn't it long since been replaced with specific entries for "G6: Deletion to make way for a page move", "G6: Unambiguously created in error or in the incorrect namespace", "G6: Redirect created by moving away from a title that was obviously unintended", and "G6: Template orphaned after TFD"?Anomie⚔19:36, 10 September 2025 (UTC)[reply]
I wanted to ask first if there was a reason no other admin had done so, and to see if anyone wanted to oppose. If no one has by sometime next week, I'll probably go ahead.Anomie⚔13:49, 12 September 2025 (UTC)[reply]
Noticed the change. I'd still include a general G6 option on the menu, just changed to "Uncontroversial technical maintenance" or "Technical maintenance" to give a clue that some technical reason is needed beyond "nobody would care, not needed". I'd also update{{db-g6}} to use the same summary and change the error dropdown to just "Obviously created in error" like the{{db-error}} summary, as the wrong namespace is still an error. ~ Jenson (SilverLocust💬)07:34, 23 September 2025 (UTC)[reply]
I'm not sure that either "Uncontroversial technical maintenance" or "Technical maintenance" really signals that a reason needs to be provided. But I won't try to stop you from making that change.Given thatWP:CSD#G6 contains "or in the incorrect namespace", I'd leave that in the dropdown. There may be someone around who considers "in error" to strictly mean "shouldn't have been created at all", excluding "creation was ok but in the wrong namespace".Anomie⚔13:41, 23 September 2025 (UTC)[reply]
I also noticed the change. There are way more than just two specific reasons for G6. If you remove the generic housekeeping option, I'll just delete without specifyingany speedy criterion. I usually specify my reason in detail anyways on most of my G6 deletions.Talk:Homo superior. –wbm1058 (talk)12:29, 26 September 2025 (UTC)[reply]
Yes, I've also just noticed it. Many of the deletions I do under this flag are not the two options that have been left behind, for example when conducting history merges. Housekeeping should be just that, uncontroversial maintenance that doesn't fall under the other specific deletion reasons, and I have no idea why we'd remove this tool from admins when it's just going to lead to less clarity.Anomie please revert your changes unless and until there is a much wider consensus for them. (We can keep the more specific instances of G6 if you like, but the general housekeeping one should still be present) — Amakuru (talk)10:42, 29 September 2025 (UTC)[reply]
Personally I had just gotten a bit tired of the mismatch between what the people who maintain this policy say inWikipedia:What G6 is not (which they've even linked from the top ofWP:CSD#G6) and the messaging in the interface which very much didn't match it. You're an admin too, you can just as well improve the messages, but I'd recommend you don't just revert the change as that would reintroduce the mismatch between the interface and policy. Even just reintroducing the "housekeeping" language wouldn't really match the current policy; you'd probably do better to say "technical deletions" instead to match what G6 has been titled as since2008.BTW, there are four specific reasons now included, not just two: "for a page move" and "in error" in the "general" section, "redirect from unintended title" in the "redirect" section, and "template orphaned after TFD" in the "template space" section (which only shows in the Template and Module namespaces). There's also nothing stopping you from typing[[WP:CSD#G6|G6]]: Whatever reason you'd like into the form without using the dropdown at all.Anomie⚔12:49, 29 September 2025 (UTC)[reply]
Per the discussion above, I have now added[[WP:CSD#G6|G6]]: Technical deletion (uncontroversial maintenance)[42]. This wording aligns with what the current text ofWP:G6 says:G6. Technical deletions. This is foruncontroversial maintenance, including: ....Mz7 (talk)23:28, 29 September 2025 (UTC)[reply]
Also, not that I necessarily disagree with the content of the essay, but I just wanted to note thatWP:What G6 is not only became linked as a hatnote toWP:G6 in January 2024 as a presumablyWP:BOLD edit[43]. I'm not sure the extent to which that essay has been vetted by the community. I know that historically some admins have been more cavalier with G6 than others, and I'm not sure if we've had any recent community-wide discussions that have explicitly endorsed the view in the essay that G6should rarely be used to delete pages that do not explicitly fall into one of the above bullet points.Mz7 (talk)23:34, 29 September 2025 (UTC)[reply]
I think we should abolish G6 entirely and split deletions to make way for moves and deletions of pages obviously created in error into their own separate criteria (and get rid ofDeleting templates orphaned as the result of a consensus at Wikipedia:Templates for discussion. entirely; that's not a speedy deletion at all, it's a deletion per a deletion discussion).* Pppery *it has begun...23:52, 29 September 2025 (UTC)[reply]
Agreed. G6 would never gain consensus if proposed today as it lumps together multiple poorly defined and only tangentially-related things. There should be separate criteria for:
Temporary deletions (e.g. for history merges)
Pages created in error and redirects created while fixing such errors
Pages in the way of a page move
Pages created temporarily as part of a page move
Pages temporarily undeleted (e.g. for a deletion review)
Anything else that meets the normal NEWCSD requirements.
I came acrossWikipedia:Redirects for discussion/Log/2025 September 12#Rotarian misogyny, an RfD that was closed as speedy delete via G8. The deleted redirect was an{{R to section}} and the target section has been removed due to NPOV violationshere. By the literal letter of G8 this is wrong since G8 deletes[r]edirects totarget pages that never existed or were deleted (emphasis mine), but since G8 uses the phrase[e]xamples include, but are not limited to which makes it open-ended, one could decide to extend the criterion to[r]edirects totargets that never existed or were deleted which would arguably make the deletion correct.Warudo (talk)15:29, 13 September 2025 (UTC)[reply]
I would say in most cases no. G8 is a broad criterion, and the list of use cases it gives is explicitly non-exhaustive, so I don't want to entirely preclude the possibility there could be a case where a redirect is truly "dependent" on the section it redirects to. But usually that won't be the case, because redirects without mention are not categorically prohibited (even if they're rather disfavored), and there may be relevant content elsewhere in the article or on another page. --Tamzin[cetacean needed](they|xe|🤷)16:14, 13 September 2025 (UTC)[reply]
Definitely not. Theonly time G8 should be used with a redirect is if the target page does not exist (i.e. it is currently a redlink) and the criterion really ought to be updated to reflect this. Sections get renamed, moved, merged and split all the time so the target section not existing means does not reliably imply anything about what should happen to the redirect (i.e. expanding G8 to included redirects to sections of pages that do exist would be a gratuitous failure of NEWCSD point 2). In the linked case, speedy deletion as vandalism or an attack pagemight have been justifiable (I haven't looked in enough detail to be certain) but G8 was unambiguously incorrect.Thryduulf (talk)16:56, 13 September 2025 (UTC)[reply]
G8 obviously does not apply in this case. Even if its listed criteria are not exhaustive, it is only for "pages dependent on a non-existent or deleted page". Sections are not pages, and removing them from their articles does not automatically make{{R to section}}s useless, as they can be retargeted to other valid articles and sections if they exist (Sexism on Rotary International (not "Misogyny on...") would have been a marginally acceptable redirect tothis section). If it were not for the fact the term "Rotarian misogyny" doesn't appear anywhere else, I would have considered raising objections to the speedy.Dsuke1998AEOS (talk)19:27, 13 September 2025 (UTC)[reply]
Let's imagine we added this criterion right now and consequently deleted thousands of redirects: Most of those deletions would be really bad and much worse than doing nothing.—Alalch E.00:36, 14 September 2025 (UTC)[reply]
Most certainly not. In this situation, G8 is strictly for redirects that outright do not work — they lead you to a page that producesMediaWiki:noarticletext. R-to-section still takes you to the target article, and whether it's acceptable without the section in question must be judged by consensus, not by speedy deletion.Nyttend (talk)20:41, 26 September 2025 (UTC)[reply]
Since we have a bunch of others weighing in here: IMO Tamzin has it right, most likely the redirect can be targeted to a different section or to the article as a whole rather than being deleted. And I particularly note that the absolutist "G8 is 100% only for broken redirects, there's never anything else that 'dependent on' could mean" position has proven to be a small minority in other contexts.Anomie⚔21:54, 26 September 2025 (UTC)[reply]
Remember: it's 1,000 views over the span of 3-6 weeks, which translates to an average of 24 to 48 views a day (for 6 weeks and 3 weeks respectively).
But I don't really care what number we use. It's just that unlike the intentionally vague term "recent", I don't see the benefit of being vague about pageviews.Green Montanan (talk)16:01, 21 September 2025 (UTC)[reply]
R3 is already one of our weirdest criteria. As written, it'll always be more of an "I know it when I see it" than a bright line. There's two obvious alternatives to that—abolish it, or define it much more strictly—but the community's never seemed much interested in that, I think because the "I know it when I see it" approach, anathema toWP:NEWCSD as it may be, in practice works pretty well. Admins delete weird redirects that'll never survive RfD, and in the rare cases where someone challenges that, RfD tends to affirm the admins' judgment. Given all that, I'm not sure how important it is to define this one term that only appears in a footnote meant to explain another vague term, as far as R3 is concerned. It may be more relevant for A10 but I don't know; is there an indication that the lack of an exact number is causing problems? --Tamzin[cetacean needed](they|xe|🤷)16:07, 21 September 2025 (UTC)[reply]
I don't really agree with this take. R3 is misused a lot, especially by editors who ignore the "recently created" rule and delete redirects that are several years old. For example,Memphis (daughter of Epaphus) was 4 years old (actually 7 years old since it was an R from move from even earlier) when it was deleted with R3. Sure, I don't think it would survive RfD so I didn't contest the deletion but it was still wrong to use R3. At best it's an IAR deletion. Perhaps we should consider either a) abolishing R3, b) rigorously defining "recently created" with abright-line rule or c) getting rid of "recently created" entirely.
On a related note, I recently contested an R3 where an admin considered an{{R from sort name}} redirect implausible enough to R3. That one is at RfD as of writing this so it may end up deleted anyway. Still, I believe that the R3 was wrong because the existence of the rcat shows that sort names are not considered implausible. In short, R3 does not workpretty well, it's just that redirects aren't that visible so people don't notice when they are incorrectly deleted.Warudo (talk)12:09, 22 September 2025 (UTC)[reply]
If an administrator is interpreting a four-year-old redirect as being "recently created" (let alone seven!), then they're not going to be deterred by an explicit numeric limit either. The problem isn't the criterion. The problem is the people. —Cryptic16:37, 22 September 2025 (UTC)[reply]
Unless we write an edit filter warning admins about too old deletions, which would be easier to do with an objective definition. I actually proposed something like this atWP:EFR once and it didn't go anywhere.* Pppery *it has begun...17:21, 22 September 2025 (UTC)[reply]
I don't necessarily mind putting some numbers on this, but this particular number is off by a couple of orders of magnitude.
Looking at the table inWP:VIEWSSTATS, only 20% of Wikipedia articles get 1,000 page views inan entire year. It doesn't make any sense to say that "very few page views" is more page views than 80% of our articles will get in a whole year, especially when you remember that they're supposed to have achieved this feat of popularity in (at most) "about 3–4 months".
The median Wikipedia article gets a total of about 20–30 page views in 3–4 months. I don't think that the median is "very few page views"; that's better described as "the median page views" or "the normal number of page views". It drops off steeply from there, though: a solid third of our least popular articles get no more than three (3) page views in 3–4 months.
But: newly created pages (including redirects) get checked by Wikipedia editors, and that process alone can generate a one-time bump of 25–50 page views (though probably not so many for redirects), especially within the first 24 hours.
Perhaps we could use these stats to find a number that is appropriate. I'd suggest maybe "averaging of one per day" as being high enough that it can no longer be considered "very few page views".WhatamIdoing (talk)16:06, 26 September 2025 (UTC)[reply]
New speedy deletion criteria "C5. Categories populated by a single page"
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.
I propose we add this new criterion, which, like its title says, are categories populated by a single page. However, it does not apply to categories that arecurrently populated with a single page, but there are other pages that fall in the category and should have the category. Yes, I read the page and in summary it has apotential and probably should be discussed.
Objective:✓ Potentially possible to tell whether a page meets this, but we might need to discuss it
Uncontestable:X Need to discuss this
Frequent:X Need to find out how frequently this occurs, maybe not
Nonredundant:X Maybe we should use this instead ofWP:CFD, but we should discuss whether it is frequent and uncontestable
So, to be clear: You think it would be preferable to delete, say,Category:Ivorian women mathematicians and have an inconsistent category structure than to keep it and have a consistent structure with some one-element categories? That seems...not uncontroversial to me. Speedy deletion should be for things that are uncontroversial. —David Eppstein (talk)02:32, 30 September 2025 (UTC)[reply]
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.
RfC: Replacing U5 with a primarily procedural mechanism
When discussion hasended, remove this tag and it will be removed from the list. If this page is on additional lists, they will be noted below.
ShouldCSD U5 be 1) repealed and replaced with a combination of 2) procedural deletion of non-contributors' user subpages after six months of no edits, 3) a narrower criterion for off-topic content that has escaped deletion under (2), 4) formalizing the practice of moving drafts off of top-level userpages, and 5) allowing editors to blank userpages that would be eligible for speedy deletion under (3) if they were subpages? --Tamzin[cetacean needed](they|xe|🤷)10:59, 30 September 2025 (UTC)[reply]
The exact details of this proposal are:
Repeal CSD U5.Text here added by Nyttend.Pages in userspace consisting of writings, information, discussions, or activities not closely related to Wikipedia's goals, where the owner has made few or no edits outside of user pages, except for plausible drafts and pages adhering to Wikipedia:User pages § What may I have in my user pages? It applies regardless of the age of the page in question.
Enact CSD U6, "Abandoned user subpages of non-contributors":
User subpages of users who have made few or no edits outside of user space, which have not been edited by a human in at least six months, excluding.js pages,.css pages, andWikipedia Books.[1] Promising drafts may be moved todraftspace by any editor as an alternative to deletion.[2]
Enact CSD U7
Original proposal, "Excessively unrelated or grossly improper non-draft subpages by non-contributors":
Alternate proposal added 02:13, 4 October 2025 (UTC), "Excessively unrelated non-draft subpages by non-contributors":
User subpages of users who have made few or no edits outside of user space, were created more than six months ago, could not be interpreted as draft articles (even very bad ones), and consist entirely of:
Creative or persuasive writing unrelated to Wikipedia (e.g.fan fictions or political essays)
InWP:UPYES, remove "usually" fromWork in progress or material that you may come back to in future (usually on subpages) and add a third sub-bullet to it:
Drafting on a top-level userpage is confusing. Any editor may move a draft away from a top-level userpage, either to a subpage or to thedraft namespace, replacing it with{{draftified userpage}}. If the user page's owner reverts such a move, this is taken as a statement that their userpage shouldnot be viewed as a draft for the purposes of assessing compliance with this guideline, which may make it qualify asexcessive unrelated content.
For users with few or no edits outside user space, excessively unrelated or grossly improper[meta 1] subpages more than six months old may be deleted underspeedy deletion criterion U7; if a top-level userpages would be eligible for deletion under that criterion if it were a subpage, it may be blanked by any editor. Note that this excludes drafts, which should be handled as describedabove.
^Wikipedia Books are a legacy feature and are slated to eventually be converted into reading lists. Seephab:T394563.
^In such a case, the act of moving the page to draftspace counts as a human edit that restarts theG13 eligibility clock.
^Note that merely writing about one's personal or professional life isnot a violation of that section unlessinappropriate or excessive orextensive[ly] self-promotional.
^ab"Lengthy" is not precisely defined, but note that it is not uncommon for established Wikipedians' userpages to have several paragraphs about their personal and professional lives. Less leeway may be given on length for content that veers closer to being promotional.
^A link indirectly providing some way to give the user money (e.g. a YouTube channel that enables paid subscriptions, or a personal website with a donation button) does not automatically make itprimarily commercial.
Meta notes
^The words "or grossly improper" will be deleted if 3B passes over 3A.
Yes. Thestatus quo of U5 is the worst of both worlds. On the one hand, its vagueness and subjectivity leave editorsfar too willing to tag and delete a wide variety of valid userspace pages, including full-fledged drafts and innocent self-introductions. In my experience as a patrolling admin, as currently stands the average U5 nomination is against policy (although often still deletable under other criteria, especiallyG11). On the other hand, it sucks up patroller and admin time leafing through pages that are largely harmless individually but dangerous in aggregate, while still missing lots of instances of that that just fall through the cracks. Few editors police old userspace subpages, meaning that even if a page could be deleted under a G-series criterion, if it survives tagging in the first few days after creation it may well never be dealt with, even if it contains BLP violations, SEO spam, etc. By having no procedural means of deletion in userspace, but relying on a vague catch-all I-don't-like-it criterion, we essentiallyguarantee negative interactions for potential constructive new users, while hosting a large enough swath of unmaintained userspace content that it's very difficult to catch actual bad actors who didn't get caught immediately.The logical solution to this is to make the deletion of unmaintained pages in non-contributors' userspace procedural, the same as it is for unmaintained drafts in draftspace.This means that the vast majority of U5 cruft will be deleted without anyone needing to assess it on the merits. U5 content is not particularly time-sensitive to delete; anything that needs to be deletednow, not six months from now, ought to be deletable under another criterion like G11 orG3. The two loopholes this would leave are where a subpage is ineligible due to being actively maintained but is completely unsuitable for Wikipedia, which the proposed U7 (preserving a stricter version of the existing U5 logic) is there to address, and where the content is on a top-level userpage, which the proposed UPNOT/UPYES changes are there for. If top-level userpage content doesn't fall under any other criterion, there shouldn't be any harm in leaving it in a page history; and if a user repeatedly restores content to their userpage that is clearly unsuitable, well, that's something they're already able to do by repeatedly recreating a deleted userpage, and in both cases this becomes a user conduct issue suitable forAIV orAN/I. The remaining cases within the current U5 scope would be borderline calls, like a page that is ambiguously either a draft or an overlongcurriculum vitae, but is being maintained and so won't get deleted procedurally; those cases are inherently subjective and belong at MfD.The other concern is, of course, deletion of useful content. I don't think this proposal would increase the rate of that, and if anything will decrease it. Currently, in the rare cases where a non-contributor writes a viable userspace draft and then abandons it, the default outcome is that it stagnates forever. Under this system, there's actually a mechanism for someone to eventually put eyes on the page and potentially move the page to draftspace, where it's moderately more likely to be noticed. When that doesn't happen, there would always be REFUND; we could renameWP:REFUND/G13 toWP:REFUND/G13 and U6. --Tamzin[cetacean needed](they|xe|🤷)10:59, 30 September 2025 (UTC)[reply]
Against enacting CSD U6, support CSD U7, wp:upyes change and wp:upnot change. The replacement of CSD U5 should allow the user to experiment until they feel like they are ready to create a draft. Failing to allow that would stifle the growth of new users, which is not helpful in the long term for wikipedia.Snævar (talk)12:37, 30 September 2025 (UTC)[reply]
I'm confused. U6 (unlike U7)does let the user experiment as long as they like - it's based on the time of the most recent edit, not the creating one. Even if they take a six-month break mid-experiment and the page is deleted, they can get it back procedurally atWP:REFUND, and I expect the standard deletion log text to point there the same way G13's does. And it'sworlds better than what we have with U5-as-practiced. —Cryptic14:00, 30 September 2025 (UTC)[reply]
U6 lets more stuff get deleted. You wrote yourself in the discussion section that newbies rarely participate in deletion discussions, so why would they ask for an undeletion?
It is better just to delete the cruft and leave the testedits, even though they could not easily get it undeleted. It can also happen, if both U6 and U7 exist, that the deletion nominator chooses U7 rather than U6, so there is not neccisarily an upside. If there is a need for a speedy due to the age of the page, there is always G13.Snævar (talk)16:42, 30 September 2025 (UTC)[reply]
To the first, that's fair. In response, I'll point at the current revision ofWP:REFUND which - like that page has been for many years now - is overwhelmingly requests for restorations of G13s. It's almost certain that the reason for that is that the default deletion log set by{{db-g13}} and the admin deletion dropdown menu item for G13 both contain instructions to go there to get the page back; so do the user talk page templates warning of impending G13. The other criteria just rely on the pink "This page was recently deleted" box saying to talk to the deleting admin, who may or may not say that the next step after they refuse is a formal appeal atWP:Deletion review. (And the pink box doesn't even show up if you're not logged in and it's been more than a couple days since the deletion.)To the second, G13 doesn't apply to user pages unless they've been submitted to AFC. And if that was done at all correctly, it'd mean that the pages were at least vaguely draftlike and so exempt from U7 anyway. —Cryptic17:14, 30 September 2025 (UTC)[reply]
I can get on board with this. U5 doesn't meet the objective criterion anymore, and I'm not sure it ever did. The rest of the changes just follow obviously.Stifle (talk)12:58, 30 September 2025 (UTC)[reply]
Yes. From my past patrolling of the user namespace from recent changes and participation at MfD and, I have come to realize that many or most editors do not have a good concept of whether someone is a contributor or a non-contributor. The concept of non-contributor is just bad for human application, causing U5 to be a bad criterion for human application as a whole. Many, for example, do not understand that a completely new user creating a draft automatically and very strongly marks that user as acontributor (they were not a contributor because they had not been a Wikipedia user, then they became a user and immediately presented themselves as a contributing user, i.e. contributor). They will then tag this brand new contributor's draft for deletion instead of moving it to draftspace or similar. That user might've been right here and could have been welcomed, notified, talked to. Instead, the very first interaction they might get is an improper tagging of their first page for deletion. At the same time, many userspace creations go unseen (they are not very interesting to patrol, unless you have a morbid curiosity), leaving the actually bad pages, of which there are many, indefinitely hosted. They need to be deleted using a catch-all criterion.—Alalch E.13:18, 30 September 2025 (UTC)[reply]
The finer points are too unimportant for me relative to the major importance of the major points to affect my !vote so I'm fine filing my support partlyen blanc. I thought about the issue of incorporating guidelines via reference in policies a lot before, and it's a tough one. Ideally, the relationship between the relevant "material norms" of the guideline on user pages and U7 as the "sanction norm" would be the same as the relationship betweenWikipedia:Attack pages (policy) andWP:G10 (policy). I'm not enthusiastic about creating new policy text that is similar to existing guideline text, so I cannot make myself engage more deeply on the finer points (edit: or, maybe, I can ...), but I still strongly support the general direction and the aims. —Alalch E.16:46, 2 October 2025 (UTC)[reply]
What: Unqualified yes on #1 (repeal U5), #2 (U6: G13-like timeout on subpages), and #4 (UPYES change). Very reluctant ok on #3 (U7: old unrelated content) based on its wording rather than its idea (see discussion below), and its dependent change #5, since while flawed, they're better than what we have and fill a hole that U6 by itself would leave.Why: my specific previous commentshere about the status quo as practiced andhere about the impact on new users. —Cryptic14:10, 30 September 2025 (UTC)[reply]
I'm good with 3B (and 5 to go with it). Also would still unenthusiastically accept 3A as second choice, provided one of the suggestions to policy-fy the referenced section, or permalink it, or similar, happens. —Cryptic02:23, 4 October 2025 (UTC)[reply]
Yes per the prior discussion. My only reservation is over U7, which I feel tries to compromise between addressing two different things.Advocacy or support of grossly improper behaviors with no project benefit shouldn't have to wait 6 months to be deleted, and I am not sure about treating it the same way as routine off-topic content.ChaoticEnby (talk ·contribs)17:13, 30 September 2025 (UTC)[reply]
The inclusion of that kind of content under U7 isn't meant to preclude earlier deletion of such content if some other CSD applies. For instance, advocating genocide would fall under G3 or G10.WP:UPNOT also already gives users discretion to blank this kind of content in any user's userspace. The only reason I included this clause is because I didn't want to create a situation where a user subpage that advocated grossly improper behaviors, but was regularly updated and didn't technically count as "excessively related" (maybe it's a page glorifying Wikipedia vandalism) would escape speedy deletion despite being by most measures worse than the merely unrelated stuff. --Tamzin[cetacean needed](they|xe|🤷)17:18, 30 September 2025 (UTC)[reply]
Oh, that explanation makes a lot more sense, thanks! Still wondering, someone repeatedly updating a page supporting Wikipedia vandalism/other project-related grossly improper behaviors, and not editing anywhere else, would probably quickly get blocked asWP:NOTHERE?ChaoticEnby (talk ·contribs)17:26, 30 September 2025 (UTC)[reply]
Probably. If this proposal passes, and I were in that situation, after blocking I would probably blank the page underWP:UPNOT and let U6 catch it in 6 months. As long as there's nothing in the page history that needs to be admin-eyes-only, there's no harm in doing it that way. That said, if others share your concern here, I wouldn't be opposed to having the "6 months since creation" part only apply to the excessively unrelated pages, allowing grossly improper behaviors to be the one thing that still get handled the way U5 currently works. --Tamzin[cetacean needed](they|xe|🤷)17:35, 30 September 2025 (UTC)[reply]
While I don't have any specific examples in mind, content that actively helps vandalism (say, giving techniques for large-scale vandalism, or for evading scrutiny) and threats/defamation against editors that don't directly fall underWP:G10 (like, asking for "someone" to dox a list of users) are the two points for which I'm strongly convinced that we don't need the "6 months since creation" clause. I'm not familiar with how common they actually are, so maybe I'm overestimating the issue.ChaoticEnby (talk ·contribs)17:55, 30 September 2025 (UTC)[reply]
Support all suggestions. I would explicitly add to U7 that it does not override other possibly applicable CSD criteria, e.g. G10. This would allow an admin atWP:REFUND to point to a specific clause when declining a request just becauseIt HaSn'T bEeN 6 mOnThS yEt! — Jkudlick ⚓ (talk)17:42, 30 September 2025 (UTC)[reply]
Support killing U5 without reservation,oppose point 4,andmeh regarding all other points (as in, I think it would be an improvement if any combination of those points were put into place, but I think there are better solutions)Weak oppose all other points. My main concern is that I think it's needlessly bureaucratic to distinguish between top-level userpages and userpage subpages. Requiring a new editor to draft something as a subpage instead of their userpage feels like an unnecessary hurdle to force them to jump through, even if the page is moved on their behalf (as thatwill lead to yet more walls of blue links that the likely already-overwhelmed editor has to wade through to understand why what they did was apparently "wrong"). If a new editor sticks with it, they will very quickly learn that they used their userpage in "the wrong way" and will self-correct. If not... so what? I really don't think there's any amount of harm caused to the project but having a half-finished draft on whatever sit on a userpage instead of a subpage.
As I said on VPIL, I really want people to keep in mind what Wikipedia looks like to a new user—in Vector-2022, you get your username in bright red text with a tooltip that just says "your userpage"—so you click that and start trying to learn how to edit things. We shouldn't lay policy landmines for people who start learning how to edit or draft or whatever by doing that. (Granted, this would be improved with something likeChaotic Enby'sWikipedia:Village pump (idea lab) § Editnotice, but edit notices are still pretty often ignored by new editors, and I don't think we should solely rely on them to guide new editors to doing things the "right" way when the "wrong" way doesn't really negatively affect much at all.)Perryprog (talk)20:46, 30 September 2025 (UTC)[reply]
Changing my vote to be more negative towards all but proposal 1 on reading through them again. I more and more think we should lean into being fine with having harmless crap in userspace, while problematic stuff should continue to be dealt with via existing CSD criteria.Perryprog (talk)18:39, 3 October 2025 (UTC)[reply]
Oppose proposal 2' (enact U6) as a solution in search of a problem. Harmful pages in userspace should be deleted for the reason they are harmful, harmless pages is userspace should not be deleted.Thryduulf (talk)11:19, 1 October 2025 (UTC)[reply]
Oppose proposal 3 (enact U7) as written. I support the spirit of this, but I oppose determining what can and cannot be speedily deleted only by reference to a subjectively written guideline - it needs to be defined objectively within the wording of the speedy deletion criterion (perWP:NEWCSD point 1).Thryduulf (talk)11:19, 1 October 2025 (UTC)[reply]
Oppose proposal 5 (changes to UPNOT) at the moment, but only because this is dependent on U7 existing and I oppose the current proposal related to that. If (and only if) a relevant speedy deletion that meets all the NEWCSD requirements is enacted then this will be fine, but not until then.Thryduulf (talk)11:19, 1 October 2025 (UTC)[reply]
Conditional support all. U5 currently does not function as intended. From the (admittedly little) I've seen, it leads to a lot of incorrect tagging andbiting of new users. Overall, this proposal seems like a sensible fix. It strengthens our ability to move drafts off of userpages, where they should not be, and it will hopefully reduce hasty tagging of borderline userpages where a new editor simply talks about themselves a little too much. However, I agree with Cryptic and Thryduulf that I want the criteria for U7 to beexplicitly written out in the CSD criteria. That is my condition for supporting. I also suggest changingPromising drafts may be moved to draftspace toPromising drafts should be moved to draftspace.Toadspike[Talk]07:50, 2 October 2025 (UTC)[reply]
Updating my !vote tosupport all, butstrongly prefer 3B (explicit criteria) to 3A per my comment in the hatted section below. Thank you, Tamzin, for putting this together.Toadspike[Talk]07:24, 4 October 2025 (UTC)[reply]
How much evidence would it take to convince you? Not rhetorical, we genuinely can generate it for you. Easiest alleged problem to prove is that its written, and community-agreed-upon, definition is almost universally ignored. You were involved in the discussions creating U5, so you're very well aware of the proper meaning of "except for plausible drafts". My experience watchingCAT:U5 is that drafts - not just plausible ones, but very obvious ones - make up at least 80% or more of the pages deleted under it.As an experiment, I ran a query inUser:Cryptic/recent U5 for the 50 most recent deletions mentioning U5, with an additional check on logid so they're spread out a bit both temporally (going back to mid-July instead of just the past ten hours) and by admin (since I don't want to single out JBW, who did most of the deletions in the past few hours). I wrote up analyses for the first five and took a look at fully half; only four, possibly five (I didn't machine-translate the one not in English) of those 25 were non-drafts, so that 80%+ percentage holds. I'll finish out the analyses on Sunday, unless someone else wants to jump in first; they'd be more than welcome. —Cryptic01:35, 4 October 2025 (UTC)[reply]
How much evidence? Let’s start with just a bit. On my part, I think I have 106 logged U5 nominations, and it seems to work to me.
The proposal conflates things, and U6 reveals and extreme deletionist Immediatist attitude, and it doesn’t appear to have a valid justification. Old project-related usersubpages are to be mass deleted? No test cases?SmokeyJoe (talk)05:16, 4 October 2025 (UTC)[reply]
U6 looks to me to be an expansion, to cover also project-related content such as essays, and plausible drafts that are not good enough to Draftify.SmokeyJoe (talk)05:43, 5 October 2025 (UTC)[reply]
U6 would be broader than U5 in conceptual scope (because it is largely content-agnostic) but narrower in procedural scope and gentler in application (because it excludes top-level userpages and any subpage that's being actively maintained, as well as allowing editors more time to establish themselves as contributors). --Tamzin[cetacean needed](they|xe|🤷)06:10, 5 October 2025 (UTC)[reply]
The start of your analysis reveals bad admin decisions. That calls for admin re-education on U5 comprehension. If bad admin decisions were a reason to o repeal a CSD user RHaworth should have seen G11 repealed.SmokeyJoe (talk)05:20, 4 October 2025 (UTC)[reply]
If the overwhelming majority of deletions under a criterion are bad, this says to me that there's a problem with the criterion. Maybe not a big enough problem by itself to justify repealing or replacing it - I've written multiple times before that, while I'm confident that these deletions aren't the sort the community authorized in passing U5, they'd almost all be bad restorations too - but more than big enough a problem than can be explained away as individually bad decisions.And this is just one aspect. Do you really need it proven to you that having your userpage deleted, justified by a reason that objectively does not apply, is usually a contribution-ending experience? I can build up stats for that, too, not to mention my own firsthand experience on multiple other wikis. (But it may take a while. My free time is currently fairly minimal.) —Cryptic03:35, 5 October 2025 (UTC)[reply]
User:Cryptic, I don’t mean to appear to be demanding proof.
I am a bit surprised by Tamzin’s proposal, I am not familiar with problems that others seem familar with.
You appear to focus on the bitiness of U5 deletions. I am ready to agree. When we were writing U5, I preferred there to be some pause before deletion, but the RfC produced a strong majority against any delay. I’d prefer a U5 deletion to be more PROD-like, with a standard REFUND route.
I think I’m fairly squarely on your page, but am asking questions. If the intent of U5 was ok, but the U5 taggers and admin deleters have strayed, why would Tamzin’s more complicated criteria be an improvement?SmokeyJoe (talk)05:55, 5 October 2025 (UTC)[reply]
I don't think it's fair to call the proposed U6 more complicated than U5. It's more objective and easier to assess - enough so that candidates could be legitimately queried for - with virtually no cognitive load compared to the current U5. I expect U6 will see a lot more usage than U7 does if both pass. And U7(B) is legitimately more complex because we've alreadyseen the kind of abuses U5 allows.(On that note, only 7 of the 50 U5 deletions in my sample looked valid, and all except one of those seven would have been speedyable under other criteria too; of the 43 I considered invalid, for all but four or five it was at least partly because they were plainly attempts, of widely varying competence, at writing drafts.) —Cryptic04:04, 6 October 2025 (UTC)[reply]
I don’t think I said U6 is more complicated, did I? It’s a little broader.
Whpq below reminds us of the problem that inspired U5, U5 fixed that problem, and so now what is left if the rough edges of U5. Decreasing the cognitive load is a good idea.SmokeyJoe (talk)07:44, 6 October 2025 (UTC)[reply]
I had not seen the village pump discussion of two weeks ago. I had no idea that our old friend User:Fastily had been driven out due misapplications of U5. Is the problem that the intent of some CSDs is not reflected in simple objectivity as read by old admins?SmokeyJoe (talk)06:12, 5 October 2025 (UTC)[reply]
Also, it's occurred to me that U6' scope would include .js and .css pages, which I didn't intend, and I don't think anyone else intended, given that there are some people who create accounts just to load custom JS/CSS. So I've added a clause on that; not going to re-ping everyone because I think this should be noncontroversial. --Tamzin[cetacean needed](they|xe|🤷)06:14, 4 October 2025 (UTC)[reply]
U7 3B sounds like a good reason, but appears to me to failWP:NEWCSD#Frequent. If these are frequent, please list ten at MfD. Frequent is important because speedy deletion is a non-transparent process, and non-transparent processes are liable to corruption.SmokeyJoe (talk)06:02, 5 October 2025 (UTC)[reply]
Yes. I am fine with both 3A or 3B. I agree with nom; it adds clarity to an extremely wishy-washy criterion. The harm caused by improper userspace pages is minimal and do not necessitate urgent deletions.Catalk to me!11:14, 5 October 2025 (UTC)[reply]
Support 1 (deprecate U5) and 3B (specific subpage criteria), andOppose the rest (2,3a,4,5). For #1, if U5 isn't working, it, by definition, shouldn't be a speedy criteria. And as for 3b, as far as I can tell, it looks pretty clear. And, I agree withMZMcBride that this (3b) could apply in all of userspace. That said, I also have seen community concerns about so-called "top-level" userpages. Tagging of userpages, for example. Which is part of why the draftify template makes me feel a bit uncomfortable. And even if that guidance was implemented, it should be a talk page note, not a big sign on their userpage. I'd rather just see a separate discussion disallowing drafts on the main userpage and the main usertalkpage. And I don't know if we want just "any user" subjectively blanking userpages. Sounds like a bad idea, laying groundwork for AN/I disruptions just waiting to happen. Also, not everyone understands the concept of the "draft userspace". What criteria are we using to decide whether the content goes to a subpage or draftspace? Also, is this going to be a page move to retain the edit history or a copy/paste move? Sounds like we're starting to get into territory of trusted editors with additional permissions. As for #2, from the discussion withThe Blade of the Northern Lights, below, it sounds like it won't address the issues that U5 was created for (due to 6 months time period), which nullifies its usage. And based upon the examples in the discussion below, can probably be handled by MFD.So to sum up, while ISupport 1 and 3B, IOppose the rest (2,3a,4,5). I am favourable to the intent behind 4 and 5, but have tooppose the execution of them as written. All that said, Ido want to say kudos toTamzin and everyone else who has been, and is, working on this. I appreciate the work that's been put into this. -jc3706:55, 8 October 2025 (UTC)[reply]
Support all after some thought, with a preference for3b over 3a. As I've said before an overhaul is needed. U5 isbitey, it's frequently misapplied, and it's a source of confusion even for experienced editors. The proposals are a clear incremental improvement: the delayed timeframe addresses a lot of the biteyness, and the rest of the criteria are much more specific. That said, there is clearly a lack of community agreement over some underlying principles: specifically, whether wewant to delete material not immediately related to improving Wikipedia. I don't think we're done with this discussion.Vanamonde93 (talk)16:07, 8 October 2025 (UTC)[reply]
2+4+5 2 is most objective. 3A and 3B can be subjective (as there are times people draft things for spaces other than mainspace, for example to experiment with editing). Should note any new criteria will have to continue the numbering as U6not U7.Tamzin Can you consider this minor fix to the CSD numbering in the original proposal?Aasim (話す)03:53, 10 October 2025 (UTC)[reply]
I don't understand. This is a proposal for two new criteria. Assuming no new U-series criterion is somehow enacted while this RfC is open, then if both proposals pass, the first criterion will be numbered U6, and the second will be numbered U7. The one presumptively numbered U7 here could in theory wind up being U6, if 3A/B passes but 2 doesn't, but so far I think only one editor has !voted yes on 3 and no on 2. --Tamzin[cetacean needed](they|xe|🤷)05:47, 10 October 2025 (UTC)[reply]
Support everything I've read through the proposal and it makes perfect sense. I don't understand the concerns about harmful pages, if they are really harmful, we have other speedy deletion criteria (most obviously G3 and G10) that deal with that sort of thing.Ritchie333(talk)(cont)08:32, 10 October 2025 (UTC)[reply]
Against proposal 2, I support everything else. If I am still alive and responding on my user talk page, I expect the courtesy of whoever is interested in the drafts in my user space to reach out to me on my user talk page, rather than have the draft unceremoniously yanked out from the user space.– robertsky (talk)01:23, 15 October 2025 (UTC)[reply]
Support all proposals I have seen some huge issues with the way U5 works and after reading and considering these proposals I believe they will be an improvement.GothicGolem29 (talk)01:02, 19 October 2025 (UTC)[reply]
@ChildrenWillListen: As an SPI clerk, I'm struggling to think of a lot of situations where this would be an issue. U6 wouldn't apply to pages that are being actively maintained, or to pages whose creators have since become active outside userspace, so the only pages this would "conceal" would be abandoned pages by non-contributor accounts, who will almost always be stale for the purposes of sockblocking—i.e. probably not necessary to block, and very unlikely to be detected regardless, as with the Hasfie sock I found while putting together§ U6: Analysis of random user subpages. I suppose there's a very narrow window where like, a user creates a user subpage with lots of socky tells, then only edits in userspace but doesn't have socky tells on any of the other pages, or where they create that page, go dormant six months, and then return to start participating elsewhere, but these both seem like pretty unlikely situations, and in either case admins would still be able to see the deleted page. Another way to put it is, acrossabout 600 sockblocks I've made, andmany more I effected as a non-admin SPI clerk, I can't think of a single one I've made that would have been hindered if U6 had existed at the time. --Tamzin[cetacean needed](they|xe|🤷)19:02, 19 October 2025 (UTC)[reply]
Support all proposals,preference for 3b over 3a but 3a would still be better than no change. There is a clear problem with U5 and I think these two new criteria would be a clear improvement over the current situation. I would also support any further workshopped version of the proposals.Choucas0 🐦⬛⋅💬⋅📋14:57, 20 October 2025 (UTC)[reply]
Support all, with preference for 3b over 3a since the former is better worded. U5 is currently one of the least objective CSD which makes it problematic and the proposed changes to towards addressing it. 4 also closes the loophole of borderline U5able content on the userpage that might also double as an attempt to start a userpage. --Patar knight -chat/contributions07:28, 22 October 2025 (UTC)[reply]
Comment: The proposal is quite clear about what is suggested should be done. But we seem to lack thereasons for the proposal. Obviously, by having a proposal to change a rule in Wikipedia, as proposed here, something is wrong or not working properly, etc. But what? It might seem clear to the author of this proposal but, methinks, they should offer a detailed reasoning, hopefully with some indicative examples, before we start digging in. Cheers. -The Gnome (talk)11:59, 30 September 2025 (UTC)[reply]
One minor nitpick, I'd changefew or no edits outside of user pages tofew or no edits outside of user space. That makes it less likely someone could misinterpret the following "which have not been edited" to be attached to the no-edits clause rather than the subpages being deleted, and makes it less likely for someone towaste time arguing that user-talk page edits count as "edits outside of user pages".Anomie⚔13:16, 30 September 2025 (UTC)[reply]
These are clear improvements over what we have (which in practice is "Any page the tagger and deleting admin don't like, in the userspace of someone too new to effectively complain"). However - and I'm sorry I didn't raise this directly in the discussions either atWP:VPI orWP:Iritalk - one of the problems with U5 is how it incorporates part of theWP:User page guideline by reference; the new U7 text, unlike the versions presented before, does so even more directly, plus calls it policy instead of a guideline. That page is a lot easier to change out from under us, both procedurally and socially, thanWP:CSD itself is, and italready leads both taggers and administrators to delete material exempted both in the current U5 and inWP:CSD#Non-criteria. Seethis recent deletion review for an illustrative example; the only thing at all unusual or uncommon about that case was that the page creator somehow managed to find their way toWP:DRV. Pointing at supplementary pages for further explanation is fine, but offloading the criterion to one to this extent really isn't. —Cryptic13:40, 30 September 2025 (UTC)[reply]
On further thought, @Cryptic, I think where I come down is this: Subjective phrases in a policy, like "few or no edits", like "credible claim of significance", etc., already behave more like guidelines than policies. They lead to an inherent variation in behavior that's more aligned with the best-practices, vibes-based approach of guidelines, than the ideally more rigorous approach of policies. So while on some level I agree with your philosophical objection to incorporating a guideline by reference into a policy, I actually think that'smore concrete than if we just used the phrases "excessively unrelated" and "grossly improper" or anything like that. And adding "unambiguously" makes clear that the greater degree of discretion usually afforded under guidelines shouldn't be applied to the same degree here. I wouldn't be opposed to strengthening the existing footnote with something like "BecauseWP:UPNOT is aguideline, not a policy, only clear-cut violations of it should go toward deletion under this criterion. On any page where a reasonable editor might conclude that the amount of unrelated content is not excessive, or that the activity advocated is not grossly improper, speedy deletion under this criterion is inappropriate, andWikipedia:Miscellany for deletion should be used instead." --Tamzin[cetacean needed](they|xe|🤷)17:31, 30 September 2025 (UTC)[reply]
It's not quite a supplementary page, it's a guideline.WP:DEL-REASON#13 already references guidelines by saying "policy" thinking "policies and guidelines" and links to the User pages guideline. I agree in spirit to your objection and it isn't appropriate that CSD should point to a guideline section which starts with "Unrelated content includes,but is not limited to:" Even though the policy proposal says "unambigously" someone could still claim that an item which does not explicitly match any of the bullets is unambiguously excessively unrelated content (might as well be, but that's not the point). The proposal could make that a little stronger (a note could say that the "but is not limited to:" part does not count for the purposes of CSD). Another thing that could be done is upgrading those guideline sections to policy sections ({{Policy|type=section}}) by running a (concurrent ?) RfC for that. —Alalch E.14:40, 30 September 2025 (UTC)[reply]
CSD could also reference a fixed version via permalink and update the permalink after the guideline has changed only per a consensus to update the permalink achieved at Wikipedia talk:Speedy deletion —Alalch E.15:16, 30 September 2025 (UTC)[reply]
The guideline section could have a "hatnote" that the section is referenced in the speedy deletion policy and that WT:CSD needs to be notified about any changes. —Alalch E.21:23, 30 September 2025 (UTC)[reply]
And just for kicks, the alternative there is blocking the user for harmlessly creating disruption in the user space. Yes, juxtaposition happened, but that would be the alternative in a lot of cases that show up at AIV or elsewhere.Izno (talk)16:10, 30 September 2025 (UTC)[reply]
@Izno: Well, I'm happy to make the general case, but first, no, I don't think that's a good deletion. I don't think it even complies with U5 as written, let alone what U5 is supposed to accomplish. It was a sandbox consisting mostly/entirely(?) of content about topics that have mainspace coverage, without editorialization, promotion, etc. Significant parts of it plausibly could have been put into articles. The fact that the editor hadn't touched mainspace lowers the likelihood of that happening, but that's the exact kind of borderline case that should be decided at MfD. Cases like that make up a small percentage of U5 deletions and would not at all overwhelm that process.Now, to the general case, the answer I'd give if you linked a regular "shitty probably-autobiographical prose CV" U5, is that there isn't much reason to delay in deleting, but also not much reason to rush, and the option where we rush both takes more editor-hours and has a higher risk of biting newcomers. The pages that need to gonow are the ones that are already covered by G3, G10, G11, and G12. --Tamzin[cetacean needed](they|xe|🤷)16:20, 30 September 2025 (UTC)[reply]
In that case I think you've correctly identified one of the rare cases that currently falls under U5, wouldn't fall under U6 or U7, and should be deleted. My question then is whether these arise often enough to justify bypassing MfD, especially given that they're already the most subjective cases and, as here, may involve a mix of appropriate and inappropriate userspace material. In my experience from patrolling U5, cases like these are pretty rare; I think a few extra MfDs a week would be a small price to pay for fixing the longstanding issues with U5. --Tamzin[cetacean needed](they|xe|🤷)16:38, 30 September 2025 (UTC)[reply]
I actually don't know the answer to that question. I only ran into this because a filter (intended for other disruption) caught the user, was checking to see how useless it was, and then found another admin (Drmies) had already CSD tagged it (though I didn't check for the tagger at the time).
No-one is going to want to spend time MFDing this page. My followup question is, how does it end up being surfaced as useless? Having to spend time looking at it every time the relevant filter (which is a good filter mind you and probably shouldn't be tightened) shows up at AIV? G13 has a useful clock from last edit. Neither U6 nor U7 would, because they'd somehow need to keep track of all the sandboxes which aren't problematic (because they don't function on 6 months from last edit, nor do they have their own namespace/category to know easily that they can be deleted).
NB I don't think on their face having criteria which function as "it has 6 months to survive utility check in our user space namespace" are bad criteria. I'm also just pretty sure tossing the bathwater here isn't a win. (But I also have neither 1: looked at none of your stats, which I suppose I might differ now on percentage given how there was disagreement here, nor 2: regularly patrol speedy deletion in general.)Izno (talk)16:47, 30 September 2025 (UTC)[reply]
Not sure I follow your point on clocks. U7 is from creation date, but U6 is from most recent edit date. The considerable majority of U6s could be found by a query for userspace subpages with more than 6 months since the last edit where the user has 0 non-userspace edits. The actual wording of the criterion is slightly broader than that (excluding bot edits and allowing deletion with "few" non-userspace edits), but that query would catch most eligible pages and provide a clock as clear as G13. Slightly broader queries could find borderline cases that humans could refine, if they want. --Tamzin[cetacean needed](they|xe|🤷)16:59, 30 September 2025 (UTC)[reply]
I'm not sold on U7 (point #3). If we're deleting userspace content because it is"Excessively unrelated or grossly improper" then the timeframe of creation doesn't feel relevant. Conversely, if we're applying a time limit, it feels largely redundant to the proposed U6, with the only additional use-case being userspace pages that are older than six months but have been edited more recently than 6 months. My inclination is to liberalize it: remove the timeframe qualifier, which would allow us to nuke, for instance, fan fiction hosted on Wikipedia, or "biographies" of children written by their friends. Conversely, if the community believes content unrelated to our purpose that doesn't otherwise meet speedy criteria isn't a problem, just get rid of U7, and let U6 handle pages as they drop into range. Of course this leaves intact the issue with U5 as it exists, which is that activity outside of userspace is a license to create as much content unrelated to Wikipedia in userspace as you want. But perhaps this isn't enough of a problem to bother with.Vanamonde93 (talk)16:09, 30 September 2025 (UTC)[reply]
(Also a reply to Izno) There are also excessively terrible draftspace drafts that barely even qualify as drafts or not at all, and they are (preferably) ignored, and get deleted via G13. (Unless they meet a speedy deletion criterion as attack pages, vandalism etc., but even when they're seen by AfC people they just process them within AfC and ignore CSD, meaning they are extra unlikely to be tagged) When has ever Wikipedia had a problem because of this practice of delaying deleting that which needs to be deleted? I don't think there's a tangible benefit gained relative to deleting the page in 6 months. That leaves room for new users and "my first page" creations to breathe, reducing the number of bad first/early interactions. —Alalch E.16:13, 30 September 2025 (UTC)[reply]
which is that activity outside of userspace is a license to create as much content unrelated to Wikipedia in userspace as you want Well, until someone decides to take your unrelated content to MFD instead of just slapping a CSD tag on it.Anomie⚔18:21, 30 September 2025 (UTC)[reply]
From my perspective the cost/benefit ratio at MfD is rarely favorable, and it becomes a worthwhile exercise only beyond a certain degree of disruptiveness. To that end I support liberalizing U7, which as written I don't believe would achieve much.Vanamonde93 (talk)20:09, 30 September 2025 (UTC)[reply]
Question –Tamzin: How will "few edits outside of user space" be defined? Will it be a fixed number or range of total edits, a percentage of total edits, or just left up to the whim of the reviewing admin? I agree with the overall concept, but I would feel more comfortable knowing there is an actual threshold. — Jkudlick ⚓ (talk)16:30, 30 September 2025 (UTC)[reply]
Yes. In the VPIL thread we played around with a few ways to improve this wording, but there wasn't any solution that I was confident would get consensus, so I stuck with the existing phrasing. I'm certainly not opposed to improvements on that. --Tamzin[cetacean needed](they|xe|🤷)16:35, 30 September 2025 (UTC)[reply]
@Pppery: Thank you. I was unaware that there had already been a consensus reached there. I'm happy with that.
@Tamzin: I really don't see a way of improving that short of stating "fewer thanx edits outside of user space", but that just sets a hard line and doesn't allow for admin discretion. I'm placing my !vote above. — Jkudlick ⚓ (talk)17:27, 30 September 2025 (UTC)[reply]
HelloTamzin, thanks for the ping. These proposals are confusing to me because they make reference to subpages but that isn't what U5 is typically used on, it's used on user pages. Could U6 or U7 be applied toUser:Brajrasprakash? --MZMcBride (talk)06:46, 1 October 2025 (UTC)[reply]
I don't have exact numbers for this, but my sense is U5 is used about equally on top-level userpages and subpages (particularly sandboxen). Regardless, neither U6 nor U7 could be applied to the page you link, although then again, neither can U5, since the page is a plausible (if non-English) draft. If the page were, rather, some kind of excessive unrelated content, then under this proposal any editor could blank it under the same logic as U7. --Tamzin[cetacean needed](they|xe|🤷)06:57, 1 October 2025 (UTC)[reply]
What you're describing is not accurate, we have historically applied U5 to pages like this for years. Pages like this are why U5 was implemented. --MZMcBride (talk)16:38, 1 October 2025 (UTC)[reply]
NeitherWP:G10 nor its more flexible cousinWP:BLPDELETE authorizes deletion merely for claims being unsourced. The onlynegative unsourced information on that page appears to be that it says "Seymour also got sober", which implies Seymour was previously not sober. It would be a valid act of BLP enforcement to remove that sentence, and I don't think anyone would in practice complain if someone just blanked the page. The broader question of potential impersonation in userspace, especially by non-contributors, and the possibility of seeming self-disclosures actually being malicious, is a good discussion to have, although I don't know if this RfC is the place for it, since this content is as much allowed under U5 as it would be under this proposal. --Tamzin[cetacean needed](they|xe|🤷)07:03, 1 October 2025 (UTC)[reply]
I have been demonstrating for a few years now that “unsourced BLP” always results in deletion after 7 days. BLPPROD should be broadened to all namespaces.SmokeyJoe (talk)08:41, 4 October 2025 (UTC)[reply]
For a bit of background on U5 (confusingly the proposal at the time was for U4), while the idea had been knocking around for a while it was a single issue that really pushed it over the threshold. At the time of implementation there was ahuge problem with people abusing userpages and user subpages to play fantasy online reality TV and game shows, a search of the MfD archives in 2013 and 2014 will give you a sense of how out of control this had gotten. Even with U5 it took until early 2017 to fully tamp this down, and I don't think anyone wants to see a resurgence. Allowing these pages to sit for 6 months would not only be delaying the inevitable, it'd actively encourage people to resume doing this since it'd get basically a free 6 months of webhosting. If U6 and U7 can address this while eliminating any vagueness in U5, I'm good with it. Pinging@Whpq: who was especially involved in that cleanup and may have some useful feedback.The Blade of the Northern Lights (話して下さい)17:26, 1 October 2025 (UTC)[reply]
What's the problem with 6 months of webhosting if it gets deleted anyway after 6 months? We're going through the same thing with G13 for drafts and it's working fine. Drafts can also be web hosting. This will actually make many pages eligible for summary deletion and will clean up userspace long-term. —Alalch E.21:53, 1 October 2025 (UTC)[reply]
Whpq and@Beeblebrox: can attest to this more than me, but giving these accounts 6 months would be an active invitation to start abusing Wikipedia userspace as stated above. These fantasy shows only last for a couple months, so by the 6 month mark they've already gotten their free bit of webhosting; sites like Tengaged were actively encouraging users there to create Wikipedia accounts for this purpose. Furthermore, these accountsnever do anything else, these pages were without exception deleted at MfD, and before U5 had become such a huge drag it was disrupting the normal function of MfD. This is not only abuse in itself, it makes it more difficult to find more serious userspace problems. Look at how many MfD entries there are in 2013/2014 forfantasy online game andfake game show to get a sense of how big an issue it was.The Blade of the Northern Lights (話して下さい)22:02, 1 October 2025 (UTC)[reply]
I'm aware of these accounts and have tagged a number of such pages for speedy deletion (as an aside, some of them can be tagged for G3 as hoaxes). The thing for me is: "6 months of free web hosting and then the pages are deleted: Okay?" There is not going to be a disruption of the normal function of MfD because MfD will reject such nominations on principle since the pages are eligible for speedy deletion and bringing them up for discussion is a waste of time (I can attest that this is how it works for drafts currently). They're going to get deleted. So 6 months of abuse and they get deleted. Doesn't sound so attractive. But we are only comparing this to: A shorter period of abuse and they get deleted + an unlimited period abuse as they do not get deleted because they are not spotted. This way all such pages will ultimately be deleted. A semi-automated process similar to G13 will take care of them. I think that more pages will be deleted actually. —Alalch E.22:49, 1 October 2025 (UTC)[reply]
There was a huge number of user pages created solely for the purpose of hosting fantasy versions of Survivor and Big Brother. There were some other shows as well, but these were the most common with Sims Big Brother being very common. These pages were only useful to the creator and game players for the duration of the game they were hosting as they were used to keep track of who was being voted out in each round (so maybe a couple of weeks). The creation of U5 allowed these type of pages to be deleted right away. Attempting to hot on Wikipedia became no longer viable as the pages didn't stay up long enough to host a game, and presumably, these people went off to find other sites to host their game (Fandom most likely). I think the question we would need to answer is "Would webhosting these games become attractive if the time period is relaxed?"
I think one of the reasons why Wikipedia was chosen as a web host for these games way back when is that we have and entire infrastructure of templates to track the results of these shows built up from the mainspace articles for these shows. It's been a decade. I would imagine if they have decamped to Fandom, they have templates that have been built to use there that support their games. This is rather speculative but my gut feeling is that we wouldn't see a big surge of fantasy games hosting here if the timeline for deletion were changed due to the inertia of being established on some other platform. --Whpq (talk)00:23, 2 October 2025 (UTC)[reply]
And if the problemdid reoccur we could institute a speedy deletion criterion that objectively targeted that specific problem without any collateral damage.Thryduulf (talk)02:47, 2 October 2025 (UTC)[reply]
Given that this problem is very well-documented, I suppose that if it does flare up again 1. people will bring these pages to MfD (since theyshouldn't be allowed to linger for 6 months) and the community will immediately see an uptick,and 2. given the history, if it gets to be anything like what it once was people will expedite putting together and enforcing a specific criterion to handle it. That basically resolves my concern regarding this issue.The Blade of the Northern Lights (話して下さい)05:29, 2 October 2025 (UTC)[reply]
The following discussion has been closed.Please do not modify it.
Cryptic,Thryduulf, andToadspike have all objected to incorporating parts ofWP:UPNOT into U7 by reference. I've already offered by defense of this above, and I think we all at least agree that this is still less vague than thestatus quo, but I do see the criticism, and so would like to offer an alternative. This is based on my initial draft of this proposal at VPIL, with some modifications based on discussion in this thread, and copying some verbiage directly from UPNOT. This also breaks "grossly improper" out of the 6-month clause as discussed withChaotic Enby (although that could be done under the currently proposed U7 too) and addressesMZMcBride's point about BLP issues, and also makes both of those apply to top-level userpages too, since the "admin-eyes-only" argument may apply in those cases.
Userspace pages of users who have made few or no edits outside of user space, which either:
^"Acts of violence" includes all forms of violence but doesnot include mere statements of support for controversial groups or regimes that some may interpret as an encouragement of violence.
^SeeWP:UPNOT regarding handling of similar material on top-level user pages.
^ab"Lengthy" is not strictly defined, but note that it is not uncommon for established Wikipedians' userpages to have several paragraphs about their personal and professional lives. Less leeway may be given on length for content that veers closer to being promotional.
^A link indirectly providing some way to give the user money (e.g. a YouTube channel that enables paid subscriptions, or a personal website with a donation button) does not automatically make itprimarily commercial.
What do y'all think of that? If this seems like a better approach (even just in relative terms, if you support both / oppose both), I will incorporate it into the proposal above as an alternate U7, and send necessary pings to those who have already !voted. Personally I think both approaches have their strengths and weaknesses; if I !voted right now it would be with equal preference. --Tamzin[cetacean needed](they|xe|🤷)10:30, 2 October 2025 (UTC)[reply]
Thanks @Tamzin. I much prefer this version (though for the purpose of achieving a consensus I would also weakly support the original point 3). It is verbose, but as the main challenge with U5 is it's vagueness, I view the detail as a benefit.Toadspike[Talk]11:42, 2 October 2025 (UTC)[reply]
I do prefer this version, with the exception of point B. A credible claim of significance isn't needed in draftspace, even for autobiographies (Wikipedia:Drafts are not checked for notability or sanity), and I don't think it should be needed for userspace either. From what I understand, these would be handled by C.iii, and the more blatantly problematic ones can already be dealt with throughWP:G10,WP:BLPDEL, or Oversight – I don't think we need to codify a separate userspace-only criterion.ChaoticEnby (talk ·contribs)13:56, 2 October 2025 (UTC)[reply]
@Chaotic Enby: I guess the situation I have in mind is like "John Smith is a math teacher at Greenacre Middle School. He thinks abortion should be illegal." That's not a G10, I don't think it quite meets the letter of BLPDEL, and it's probably not OSable. And yet, it's definitely an unsourced contentious statement, and probably shouldn't stick around for six months. I don't know, I realize I'm kind of arguing from both sides of this—occupational hazard of trying to find a compromise position that keeps as many people happy as possible—but it seemed worth it to me to pull this scenario out of the age and non-draft requirements. Will await others' thoughts before deciding whether to include it in any formally proposed alternative U7. --Tamzin[cetacean needed](they|xe|🤷)16:14, 2 October 2025 (UTC)[reply]
What I'm thinking is that, if "John Smith is a math teacher at Greenacre Middle School. He thinks abortion should be illegal." doesn't fall under G10 and wouldn't be speedily deleted in another namespace (say, draftspace) without waiting for six months, then there shouldn't be a fast-tracked deletion process exclusively for userspace. I completely understand the reasoning behind it, but I feel like it should be in a wider discussion about expanding G10's scope rather than making it part of U7.ChaoticEnby (talk ·contribs)16:32, 2 October 2025 (UTC)[reply]
I wouldn't be opposed to yoinking that verbiage out of U7 to become G16, and/or expanding BLPDEL to include it, given what Alalch is saying below about this being a frequent issue at MfD, but that's not mutually exclusive with passing this as U7B now (as part of a proposal that's already got fingers in more policy sections than is ideal). --Tamzin[cetacean needed](they|xe|🤷)16:58, 2 October 2025 (UTC)[reply]
I'm fine with B based on the "credible claim of significance standard" and B is sufficiently frequent because unsourced BLP content about entirely private people is not rare at MfD and is always deleted because of BLP reasons -- even in draftspace,WP:NDRAFT notwithstanding. Said standard is an objective and uncontestable standard that is already used to specify speedy deletion criteria, so it is a tried and true thing. About whether B includes the user themselves: I agree with Pppery that the answer should be no. And I would add: It should not be assumed that the content is about the user if it is not explicitly about the user: It should be assumed that it isnot about the user. —Alalch E.16:34, 2 October 2025 (UTC)[reply]
In that case (especially if they are always deleted in draftspace too), it should probably be an expansion of G10, which I would support, rather than a specific userpage criterion. To clarify, if these are always deleted already, I dosupport the proposed U7 as a much better alternative to the status quo, although I do believe that it gives us the opportunity to carve out U7B into a wider criterion in a followup discussion.ChaoticEnby (talk ·contribs)16:40, 2 October 2025 (UTC)[reply]
It isn't fair to lump it with G10 andwp:Attack pages because no one is being attacked. The content is bad and strongly unwanted, but it just isn't natural to label it as attack page content. That would bear too harshly on the erring users. —Alalch E.16:48, 2 October 2025 (UTC)[reply]
To clarify, what I meant was expanding G10 to include non-attack unsourced BLP statements about private people, not labeling these as "attack pages" under the current G10 definition.ChaoticEnby (talk ·contribs)16:52, 2 October 2025 (UTC)[reply]
As an arm of a G criterion, that would be redundant toWP:A7 in article space, and it's a little iffy to specify an arm of a G criterion as effectively only a U criterion —Alalch E.16:59, 2 October 2025 (UTC)[reply]
My idea was for it to apply at least to both draftspace and userspace, which a U criterion can't really do. But, as I said above, I support it being U7B for now, as much better than nothing at all.ChaoticEnby (talk ·contribs)17:08, 2 October 2025 (UTC)[reply]
Yeah, I understand. For future reference, here are some"unsourced BLP" MfDs. Some pages may contain a credible claim of significance and most are the user writing about themselves, but still, it can be seen that even drafts are routinely deleted and are not left to G13, for BLP reasons, unlike drafts about other topics (deleting them faces stiff resistance, in contrast). I'd probably support a G16 derived from U7B and A7. —Alalch E.17:21, 2 October 2025 (UTC)[reply]
This is gettingreally complex. As it stands, A and B should probably be a different top-level criterion than C, due to the latter's age and subpage restrictions (not to mention having itsown sub-subcriteria).Alternately, maybe we could combine them all into a shorter minimum page age and make C apply to top-level pages too? I can't think of anything that would fall under A that wouldn't be speedyable as G3 vandalism (take a moment to read the first sentence ofWP:Vandalism if you haven't for a while). And the rule of thumb I used for U5-tagged pages consisting mostly of links to social media was whether the page was a month old yet with the user still having no mainspace edits. Six months seems overly generous, and one might be short enough to delete the stuff in B that isn't already G10able.Or I'd be fine with one of Alalch E.'s earlier suggestions of permalinking to WP:UP instead of just linking whatever its current version is, or tagging it as policy-section. —Cryptic01:09, 3 October 2025 (UTC)[reply]
Hmm.@Toadspike,Chaotic Enby,Alalch E., andVanamonde93, what do y'all think? Propose with the three subcriteria, or just part C and loop back to B later as part of a potential G16 discussion (on the assumption that A already falls under G3/G10)? I'd be opposed to going shorter than six months on C, though, as part of the logic should be that you only need to consider it for pages that are escaping attention under U6. Otherwise we're inviting a wave of policing of marginal pages that would have just been U6'd a bit later anyways. --Tamzin[cetacean needed](they|xe|🤷)04:40, 3 October 2025 (UTC)[reply]
Maybe: Biography of a deceased person that mentions some living people. Has references but no links. Moderately promotional but not in a commercial way.
Maybe: Subject might be notable (I'm guessing the sources showing it wouldn't be in English), but draft would need a near-complete rewrite.
Yes: OverlookedG3 (hoax, as far as I can tell) andlikelyG5 (whichever sockmaster[46] wasHasfie).
No
So on question 2, out of 20 sampled, we have 7–8 cases where maintaining the page is doing some level of ongoing harm, including 3 which are overlooked G-series speedies. On question 3, we have 12 cases where absolutely nothing of value is lost in deleting, 4 that might have the slightest bit salvageable but probably not, and 4 that are books, which are sorta their own thing. I'd be fine with exempting books from U6, although the book feature is almost completely obsolete so it might not be worth it. --Tamzin[cetacean needed](they|xe|🤷)07:22, 4 October 2025 (UTC),ed. 07:26, 8 October 2025 (UTC)[reply]
Largely agree, but disagree with books and citations. Generally newbies do not know how to cite. They probably did learn how to do cites in school and might do something like "lorem ipsum eret.(Miller, Edward (2005). The Sun. Academic Press.)", but they are unlikely to use <ref> tags. They do need to be told something like: 'Your sandbox at X is missing citations. "You can add a citation by selecting the button at the top of theediting box."' (this sentence and "Miller, Edward..." fromWikipedia:Citing sources ). The cite image is important and is impactful on whether a newbie understands the instructions or not. Generally, it is best to assume that a newbie does not know any wikicode or UI buttons whatsoever.
I read Tamzin’s proposal not as a repeal and fresh proposals but as a re-working of U5.
I read a major criticism of recent years U5 deletions as their bitiness, especially when done wrong.
I believe that U5 as originally approved served its purpose. The need was to respond to a very large amount of blatant NOTWEBHOST violations by non-contributors, and their nominations at MfD overwhelming MfD, with SNOW delete results.
The bitiness largely comes from the term “speedy deletion”, which in plain English implies there is not point asking or arguing, it is so badly unwanted we don’t care what you have to say.
I suggest that going forwards, U5 deletions (or U6, U7 as proposed, though I suggest not creating these as named) be a PROD like deletion process, where a gentler message (“it is proposed”) is posted on a talk page, and the author is invited to respond. If there’s no response in seven days, delete. Mostly, the users won’t respond in seven days, but with an expired PROD, it is much easier to ask late than it is for an old speedy deletion.
I suggest describing the scope as “non-contributor’s old pages”. Implied is that the pages cannot be read as intended contributions, which includes even poor drafts, and left-field comments or criticisms about Wikipedia.
No way. PROD is for mainspace only. If you want to add a step of talkpage notification before csd-tagging, that might be nice, but absolutely no to prod outside of article-space. -jc3707:01, 8 October 2025 (UTC)[reply]
What is the reason for "Redirects with no space before a parenthetical disambiguation" to not be in the redirect category, named X3 instead of stuff like R5 or R6?BodhiHarp16:14, 30 September 2025 (UTC)[reply]
Because A) when newly-created, these redirects are alsoR3s; and B) when worded like this, there's a presumption (at least to some admins) that any title with an open-paren not immediately preceded with a space is speedyable, even when it's not a disambiguator, and even when not having the space iscorrect. (User talk:Cryptic#Improper disambiguation redirects for concrete, coincidentally-timed, examples.) Making it a temporary criterion makes it easier to remove it and limit the damage after all the old ones have been dealt with. —Cryptic16:27, 30 September 2025 (UTC)[reply]
Because of those examples that did not actually meet X3 because they were not disambiguators, should we clarify that it only applies to disambiguators?BodhiHarp16:42, 30 September 2025 (UTC)[reply]
...it's already right in the criterion's title. Admittedly it's not in the criterion's text, which defines it solely with two examples, followed by some disqualifiers. —Cryptic17:00, 30 September 2025 (UTC)[reply]
There are an increasing number of drafts created from the outset with a fake "decline" review template. Somewhere there's an AI that thinks it must be included. It's becoming one of the hallmarks of an AI-generated AfC draft.Draft:Shawn Hale is a recent example, including a fake review template and content warning tags (notability and refimprove) from the first edit.
I propose that if the person who creates the draft cannot even bother to remove erroneous templates, it should be considered as unreviewed AI output and qualify for G15, especially if there are other indicators of AI generation.
I have to say, "AI pre-declines its own drafts and its human meat-puppets keep that when posting it" is pretty hilarious.Anomie⚔22:18, 7 October 2025 (UTC)[reply]
Offhand, I feel like I've seen a few through my patrolling, but there is the risk that it might be a confused user with a local backup of a previous draft. To be safe, it would be best to restrict the new criterion to previously undeclined drafts with a decline template and drafts with a decline template that has been modified any previous versions. --Patar knight -chat/contributions00:39, 8 October 2025 (UTC)[reply]
The difference here, if you look at the example I gave, is that the decline template is basically content-free with no reason given. That's an indication that it isn't a previously declined draft. ~Anachronist (who / me)(talk)01:53, 8 October 2025 (UTC)[reply]
I'm inclined to disagree that it should qualify for a G15 due to the presence of a decline template alone. Copying my comment over fromWikipedia:Teahouse#Draft:Shawn Hale:
I'd say it's more a glaring hallmark that AI was used in the process, but I AGF that it doesn't always mean the content itself wasn't reviewed as required for G15. New editors are (quite understandably) unlikely to understand or change template coding – especially when the decline parameter in this case is a simple "d" between the pipelinks, e.g.{{AFC submission|d|ts=20251005}}.
I take it as a sign to be extra skeptical of the content, but imho it's not a 100% guarantee the body content itself wasn't reviewed or rewritten by a human (but it's fairly obvious when it hasn't been).Nil🥝00:45, 8 October 2025 (UTC)[reply]
The decline template can be explained by an editor restoring an old draft from backup. But I do believe egregiously incorrect template use can be a sign of unreviewed AI use, in the draft given the WikiProject talk page templates are used at the bottom inappropriately. Perhaps something that covers cases where the templates are used incorrectly and which no human could reasonably make would work?JumpytooTalk03:53, 9 October 2025 (UTC)[reply]
That would be exceedingly difficult to make objective. Especially when you remember that most good-faith new editors arrive here with (near) zero knowledge of what templates are, let alone the complexities of how they are and should be used and the syntex, etc required to actually achieve that.Thryduulf (talk)04:22, 9 October 2025 (UTC)[reply]
Based on the comments here, it seems more and more like an empty decline template should be mentioned in the paragraph following the bullet points, as a signal that this is AI output thatmight be unreviewed, and should be confirmed by other means. ~Anachronist (who / me)(talk)06:29, 9 October 2025 (UTC)[reply]
Remove Getty Images as an example fromWP:F7. That website is well known forcopyfraud (Getty is mentioned on that article twice) and an image which comes from there should be scrutinized instead of being speedy deleted. ―Howard •🌽3321:47, 22 October 2025 (UTC)[reply]
Non-free images or media from a press agency or photo agency (e.g. Associated Press, Getty Images), where the file itself is not the subject of sourced commentary, are considered an invalid claim of fair use and fail the strict requirements of Wikipedia:Non-free content criteria, and may be deleted immediately.This criterion does not apply if there are good faith concerns that the agency doesnot hold the image copyright (e.g. the image is actually in the public domain), such concerns should be discussed at [venue].
Is this a realistic concern? Getty clearly does engage in copyfraud, but are there examples of files sourced to the site being speedy deleted as copyright violations when they shouldn't have been? Every file I've seen sourced to Getty Images (or later found on Getty Images), both here and on Commons, has been a clear copyright violation. -Eureka Lott15:12, 23 October 2025 (UTC)[reply]
I'd suspect you'd find more instances of the problem on Commons than here, since public domain images here tend to get transferred there.Anomie⚔16:24, 23 October 2025 (UTC)[reply]
Getty certainly has tried to license images for which they do not hold the rights. Having acknowledged that, I suspect that much of the Getty material uploaded is copyrighted by Getty or somebody who has contacted with Getty. If a Getty image is freely licensed, we world still need proof of that licensing. For example, a US Govt image claimed by Getty but can be shown to be PD from some other source that establishes that images proper license such as Library of Congress entry.Whpq (talk)15:47, 23 October 2025 (UTC)[reply]
RFC: Updating T5 to account for parent templates that have been merged at TFD
Recently there have been a large number of merges viaWP:TFD that have resulted in unused subpages. One such example that comes to mind was{{Infobox Twitch streamer}}. Once that merge was completed, the main template was redirected, but that left the /sanbox, /testcases and /styles.css pages. A user attempted to speedy delete them, but that was rejected by an admin becauseWP:SDdoes not include "subpages of merged templates". Thus for these unused templates were forced to file a fullTFD.
Now I would argue there are other criteria that fit...
WP:G8 (Pages dependent on a non-existent or deleted page) could also be interpreted to pertain, though in this case{{Infobox Twitch streamer}} was nottechnically deleted, it was redirected.
But given that there are a number of admins who have felt in the past that these types of pages do NOT fit T5, I think we should spell out the fact that they do to make it clear. I have not heard ANYONE say that these types of pages shouldnot be deleted. It has been a purely an objection on a technicality.Zackmann (Talk to me/What I been doing)00:26, 23 October 2025 (UTC)[reply]
Oppose If the main template is deleted, then G8 applies. Otherwise the /doc, /sandbox and /testcases should be redirected to those of the target of the original template rather than deleted.* Pppery *it has begun...00:35, 23 October 2025 (UTC)[reply]
WP:DEL-REASON #10:Redundant or otherwise useless templates.
WP:TFD#REASONS #3:The template is not used, either directly or by template substitution (the latter cannot be concluded from the absence of backlinks),and has no likelihood ofbeing used. (emphasis from source).
Oppose per Pppery. In most cases redirection following a merge is going to be preferable to retain edit histories, avoid breaking links in discussions and help understand the evolution of templates. In the few cases where redirection would be harmful but no existing speedy deletion criterion applies, then TfD is more than able to handle the load.Thryduulf (talk)02:41, 23 October 2025 (UTC)[reply]
button at the top of theediting box."' (this sentence and "Miller, Edward..." fromWikipedia:Citing sources ). The cite image is important and is impactful on whether a newbie understands the instructions or not. Generally, it is best to assume that a newbie does not know any wikicode or UI buttons whatsoever.