Wikipedia:Village pump (policy)

I've heard it discussed around a fair amount, and I'm sure it's one of those 'perennial proposals' that the veterans here are going to roll their eyes and say "ugh, somebody's bringing THIS up again," but I do think it bears saying. Notability is an awful descriptor for what we're actually looking for, which is presence in sources. That's 'notedness' if anything, not 'notability', and the inevitable result is that every time you tell someone you can't accept their autobiography/company's article/article about their favourite media thing because it's 'not notable,' they get their haunches up and go on a tirade about how many awards they/the thing have won and how many cool things they/the thing have done, etc. Pretty much every mention of something being notable or not notable has to be accompanied by a mandatory disclaimer of what notability means here and how it doesn't mean what they think it does. It's a thought particularly spurred on by my deletion nomination of the article Deaglán de Bréadún, which led the man himself to post a response essentially calling me a nasty person for daring to imply that him and his career aren't notable... which, of course, is not actually what we mean, despite literally saying the words "you aren't notable enough for a Wikipedia article"

So, the obvious question is; what would we call it instead? I've heard the term "Criteria for inclusion" mentioned, which I think would be a graceful solution, since you can explain that the criteria for inclusion is presence in sources etc without ever having to use the scary word 'notability.' Whatever alternative option is presented, I do think it is seriously high time that Wikipedia take the big step of retiring the term 'notability' Athanelar (talk) 00:39, 20 February 2026 (UTC)

I agree with you. Notability is a dumb name. However, there's never going to be a consensus to change it. voorts (talk/contributions) 00:55, 20 February 2026 (UTC)

And then, what would we call all the lists of "notable" people/residents/alumni/etc.? "People/residents/alumni/etc. who meet the criteria for inclusion"? Donald Albury 01:11, 20 February 2026 (UTC)

I think you could probably still keep those; the definition there would logically run in the opposite direction, they are notable because they meet the criteria for inclusion. It's not an ideal solution, but obviously cuts down on some of the logistical challenge. Athanelar (talk) 01:41, 20 February 2026 (UTC)

Renaming notability has been an WP:PEREN issue, repeated discussed without finding any term that has a benefit over "notability" that would not be disruptive (how many P&G depend on it) but would be more descriptive. And no, "presence in sources" is an indicator of notability, but not how notability is defined. Masem (t) 01:16, 20 February 2026 (UTC)

It is undoubtedly true that it would be a lot of work changing the nomenclature all over the project if we got rid of 'notability,' but I don't think "It'd be a lot of work" is a compelling counterargument. There would no doubt be a long transitional period where lingering remains of 'notability' were still all over the place; but that doesn't have to be an issue, it's not like the change has to happen overnight. We could just say "from today on, the term 'notability' is deprecated and we prefer this new term instead" and people can change mentions of it as and when they catch it. Just today I heard about the NSPORTS 2022 RfC where the definition of notability for athletes was radically changed; that too would've had far-reaching implications on the project, but that didn't stop people from doing it just because it was a lot of work ahead of them to clean up the now non-notable athlete articles everywhere. Athanelar (talk) 01:42, 20 February 2026 (UTC)

The NSPORT change did not radically change what notability was, just eliminated a very poor presumption of notability (playing one professional game) that had led to thousands of permastubs on athletes that was a constant problem at ANI.

We've been through what the downstream impacts of changing the term notability to something else as part of past discussions (because this being PEREN) and its not as simple "from now on it will be known as..." "notability" is embedded in WP culture and in coverage of how WP works, so it would be a massive shift, so any new terms must carry a lot of massive benefit to make it worth the effort to make the change. And dozens of suggestions have been made and failed to show this. Masem (t) 04:47, 20 February 2026 (UTC)

• My 9-year-old essay's time has finally come! WP:Noted not notable. (Note: It's a very, very short essay, admittedly.) EEng 01:58, 20 February 2026 (UTC)

  We could put this in big letters on the notability page and all the spin-off pages like WP:42. Thebiguglyalien (talk) 🛸 04:24, 20 February 2026 (UTC)

  This is also a fairly elegant solution, hut unfortunately relies on people actually clicking the link and then actually reading the words on their screen. The 'headline effect' is real when it comes to WP links. Athanelar (talk) 09:23, 20 February 2026 (UTC)

  Nice essay, EEng. I especially like how the "nutshell" explanation is nearly twice as long as the essay itself  ;) —Fortuna, imperatrix 19:24, 20 February 2026 (UTC)

  EEng 20:16, 21 February 2026 (UTC)

• For reference, the last big discussion on this topic that I know of is Wikipedia talk:Notability/Archive 84 § RfC on change of name, from April 2025. isaacl (talk) 02:07, 20 February 2026 (UTC)

  Much appreciated. Starting to read through some of that, I think 'eligibility' is quite a strong contender. It's an easily-understood English word already, which perfectly encapsulates the concept being described, and also would actually allow us to broaden our definition in some useful ways (because eligibility is a catch-all term that could include not only the presence of positive indicators like strong sources, but also the absence of negative indicators like WP:NOT or criteria for speedy deletion.) A subject which is 'eligible' is one that is both suitable for inclusion and unsuitable for deletion, which 'notability' does not currently encapsulate. Athanelar (talk) 02:15, 20 February 2026 (UTC)

  No, eligibility is a terrible idea, because it implies a brightline yes/no answer. Notability is a greyscale, its why notability is based on presumptions and not a hardline test.

  The only real issue with notability is for editors encountering the term for the first time, and coming to learn that real-world definition of notability is not exactly the same as WP's definition of notability, but reading the P&G should quickly resolve that. Masem (t) 04:53, 20 February 2026 (UTC)

The only way to get the name changed? would be to propose only one alternative. GoodDay (talk) 02:10, 20 February 2026 (UTC)

I disagree, the bugbear for a lot of people seems to be whether we'll get consensus that changing the name is worth the effort. I think people first have to be disgruntled about the old name to be resolved to change it before we worry what the new name ought to be. Athanelar (talk) 02:11, 20 February 2026 (UTC)

Do you have a proposed name? GoodDay (talk) 14:37, 20 February 2026 (UTC)

No! God please no! This is a perennial issue based primarily on WP:IDONTLIKEIT. It’s too late and too entrenched to change and someone is just going to bitch about how dumb “eligibility” is down the line. A better proposal would be outright banning perennial proposals and requiring consensus to unban them before allowing them to be discussed again, since that would require more extraordinary reasoning than “I know this has been talked to death, but just me out, I swear”. Dronebogus (talk) 02:43, 20 February 2026 (UTC)

While I agree that changing it has a less-than-ideal cost-reward, this post (and the others before it) explain legitimate downsides to the current name besides preference. Also, a discussion to unban a perennial proposal would look almost identical to the proposal itself. Thebiguglyalien (talk) 🛸 04:28, 20 February 2026 (UTC)

A discussion to unban a perennial proposal would look almost identical to the proposal itself. Maybe so, but it would force proposers to go through the process twice, which would discourage most proposers from doing it at all and save everyone a lot of time. Additionally, it wouldn’t necessarily always result in the aforementioned situation— if a proposal was banned because it was a hot-button issue now, it might be uncontroversially removed from the list 10 years later after things cool off, without actually endorsing it. It would be sort of like the MediaWiki:Bad image list or a gold lock for proposals. Dronebogus (talk) 09:51, 20 February 2026 (UTC)

I totally agree with you and I was disappointed that there wasn't consensus to change the name in the aforementioned April 2025 RfC. But given the outcome of said RfC, I struggle to see the point of rehashing the discussion so soon as it's very unlikely that there will be a different outcome. Perhaps give it a year or two.  novov talk edits 05:24, 20 February 2026 (UTC)

• Coverage perhaps? Or renown? Or just noted ... I doubt it'll ever actually change though. EvergreenFir (talk) 05:42, 20 February 2026 (UTC)

  Coverage is a distinction without a difference. Renown is far more pretentious than notability. Noted is barely even a change and couldn’t be used rationally in a sentence. Dronebogus (talk) 09:53, 20 February 2026 (UTC)

• WP:Eligibility was recently suggested by Wikipedia expert Bill Beutler. SmokeyJoe (talk) 06:00, 20 February 2026 (UTC)

  Eligibility is the rename option that sucks the least, since it most accurately reflects what “notability” actually means. But “notability” still at least puts a vague idea in people’s heads (namely, is this thing/person significant in some way?) whereas “eligibility” could mean pretty much anything and could actually put a worse idea in people’s heads (namely, eligibility is whatever somebody says it is, or is some arcane ruleset known only to insiders that isn’t easily summarized). Dronebogus (talk) 09:59, 20 February 2026 (UTC)

  I still think "notedness" would be better than "eligibility", as "eligibility" seems like it should include non-WP:N things such as WP:NOT and WP:BLP. Anomie⚔ 13:50, 20 February 2026 (UTC)

  I don't see that as a negative. We are, ultimately, looking for a term that describes "eligible to be included on Wikipedia." In fact, some of the AfC decline notices literally use "your references do not demonstrate that this subject qualifies for a Wikipedia article" as a piped link to 'notability' anyway. If anything, having a more comprehensive term would be an advantage, since then you don't run into the tricky situations of 'well, we TECHNICALLY have enough information to presume this person is notable, but there's still not enough coverage to substantiate an article about them' amd so on.

  Eligibility includes what we now define as notability, but way more succinctly communicates the point of whether or not something should have an article. Athanelar (talk) 14:02, 20 February 2026 (UTC)

  Having an all-encompassing term actually referring to just one facet seems like it would make it harder to discuss that facet versus other facets. "Ok, that meets WP:ELIGIBILITY, but it's still not eligible because it doesn't meet WP:BLP1E." "Wat?" Anomie⚔ 00:01, 21 February 2026 (UTC)

  Exactly. Dronebogus (talk) 14:33, 21 February 2026 (UTC)

  “notability” still at least puts a vague idea in people’s heads: but I think that’s the problem with the term. People don’t realize they are encountering a jargon term and substitute their own meaning. I’d argue that “eligibility” is better because there’s more precedent that contextual criteria will define eligibility for a particular thing; it might cue people that they need Wikipedia-specific information. (I’d almost want to try a complete neologism that people would know they don’t know the meaning of, something like “wikifiability” or “AAOEW” (Article Allowed On En-Wiki) that they’d know they don’t know.) ~ le 🌸 valyn (talk) 07:35, 21 February 2026 (UTC)

  Anomie said it best with the “wiki-eligibility is not dictionary definition eligibility” example. Notability still hits the general vicinity of the right idea; eligibility doesn’t and has the potential to be more confusing. As for a neologism, I don’t support that either because (on top of being a solution in search of a problem like all these replacement suggestions) it just adds MORE incomprehensible jargon to Wikipedia— which is what this proposal is supposedly trying to cut back on. Dronebogus (talk) 14:39, 21 February 2026 (UTC)

  I feel like you might have misunderstood my argument. When it comes to the one-word name for this concept, I contend that "trying to cut back on" jargon is counterproductive; any one-word name for this mess of concepts is inherently jargon. Accordingly, I think there's no point trying to change to something "clearer", but it could possibly be helpful to change to something less "clear", because it could make the term into a "known unknown" instead of "something you know that isn't so". Personally, when I want to avoid jargon with newbies, I write out a whole explanatory phrase instead (eg "our criteria for a book to have an article"); I think that's the only approach that can actually effectively cut down on jargon. ~ le 🌸 valyn (talk) 02:42, 23 February 2026 (UTC)

• Notability was never a good choice of name, but we've stuck with it because of the cost of changing; it's a QWERTY vs DVORAK problem. Personally I'd quite like to call it "Citability".—S Marshall T/C 09:45, 20 February 2026 (UTC)

  I think the 'cost of changing' would quickly be outweighed if you tallied up the editor-hours required and compared it to the editor-hours that have been spent and will continue to be spent in perpetuity explaining to disgruntled would-be article creators why appearing on a list of the Top 100 Best Things doesn't confer notability. Athanelar (talk) 13:11, 20 February 2026 (UTC)

  Don't misunderstand: I personally would support most reasonable alternative names because "notability" is still (after all these years) a misleading, dramagenic, and generally awful choice of words. But there's a non-zero cost of changing and a lot of neophobia to overcome here.—S Marshall T/C 14:40, 20 February 2026 (UTC)

  Is “dramagenic” a word? Dronebogus (talk) 14:40, 21 February 2026 (UTC)

  As I discussed last April, personally I encourage everyone to focus on providing more complete explanations on the standards for having an article rather than just linking to a jargon term. The key obstacle is that the community has to want to reduce its use of jargon. isaacl (talk) 16:51, 20 February 2026 (UTC)

• It is highly unlikely that the community is going to rename "Notability", this being as noted a perennial proposal that gets enmeshed in the long and complicated history and complicated current understanding of the concept of 'notability' on en.wiki. However, a creative smaller change probably worth exploring might be to create an alternative name for WP:GNG that somehow does not include the "N". GNG is the aspect of notability that best describes "presence in sources", it is the least likely aspect of notability to get enmeshed in notability politics. I don't have a perfect suggestion offhand, but creating an alternative name for GNG is a smaller task then renaming all of notability, and would capture much of the practical benefit of a full notability rename even if that full rename never happens. CMD (talk) 14:11, 20 February 2026 (UTC)
• The community's inertia is such that a proposal to change this isn't a good use of my time or anyone else's. But I agree, and if I had my way I would want the policy not to be a near-synonym of "significant". The practical consequence I see most often is the eliding of "should we as an ambitious global encyclopedia cover this in principle" and "can we as an encyclopedia that cares about verifiability write an article about this in practice". I could go on at length, but a more prosaic name may help us a good bit, perhaps something as plain as "standard for inclusion". Vanamonde93 (talk) 17:47, 20 February 2026 (UTC)

  “Standard(s) for inclusion” is by FAR the best proposed option so far; but there are multiple “standards for inclusion” beyond notability. Dronebogus (talk) 14:46, 21 February 2026 (UTC)

  @Dronebogus: This is precisely the distinction that I want to make. To my mind there's not multiple standards for inclusion, because in our most overarching policy language, "notability" is used as a synonym for standards of inclusion. We do have multiple was of showing notability beyond GNG/SIGCOV. And the frequent use of "notable" to mean "has SIGCOV" therefore causes considerable confusion as well. Vanamonde93 (talk) 20:18, 21 February 2026 (UTC)

  "Notability" is one commonly discussed standard, but there are others such as WP:NOT and WP:BLP. Anomie⚔ 20:52, 21 February 2026 (UTC)

  @Anomie: I see those as standards for exclusion, which may require the removal of notable topics, but will never compel the inclusion of non-notable topics. Vanamonde93 (talk) 01:49, 22 February 2026 (UTC)

  Two sides of the same coin. You could just as well say that WP:N will never compel the inclusion of topics that go against WP:BLP, and so on. It all goes together to determine what's included, i.e. multiple criteria. Anomie⚔ 03:20, 22 February 2026 (UTC)

• Wikipedia jargon. We use a word with a meaning that differs from its normal English meaning. Any other word would therefore have the same issue unless we created an entirely new word like "cituated". My personal favourite is "living persons", which includes dead persons. Hawkeye7 (discuss) 18:42, 20 February 2026 (UTC)

  I don't think that this is necessarily true; "eligibility" "suitability" and "criteria for inclusion" which have all been mentioned all transparently mean what we would want them to mean; the bar a subject has to pass to get an article. This isn't an unsolvable dilemma, it's just hard work. Athanelar (talk) 19:33, 20 February 2026 (UTC)

  "Criteria for inclusion" leads us straight to one of the most common points of confusion: inclusion of an article in the encyclopaedia versus inclusion of details within an article. The criteria applied to the creation or retention of an article are not the same as those applied to the content inside it. Hawkeye7 (discuss) 20:36, 20 February 2026 (UTC)

• "Notability" really is a terrible name. Yes, you are correct. Toadspike [Talk] 18:49, 20 February 2026 (UTC)
• Yes, it is really a terrible name. The fixation that it needs to be one word is also bizarre. Neutral Point of View is not one word, Original Research is not one word, Biography of Living Persons is not one word, Article Title is not one word, etc.: so, Article Criteria, or some such. 'On Wikipedia, Article Criteria is a test . . .'; It meets the AC; it does not meet WP:AC; and done. Alanscottwalker (talk) 01:44, 21 February 2026 (UTC)

• Comment: Unpopular opinion I guess but I like the word "notability," especially when paired with "Wikipedia:Verifiability." Notability gives a lot of wiggle room but suggests there is some minimum for inclusion, and we can adjust what that is.

GeogSage ^{(⚔Chat?⚔)} 02:23, 21 February 2026 (UTC)

It’s not an unpopular opinion. I’m pretty sure the silent majority either likes it or has no strong opinion on it. Otherwise we would have changed it by now. Dronebogus (talk) 14:50, 21 February 2026 (UTC)

• I think some people are confusing the aim and the criteria. We really do want to write articles on topics that are notable according to its everyday meaning (that's the aim), but to achieve that in practice we have to make guidelines for notability that editors are able to follow and agree with each other about (that's the criteria). So my opinion is that "notability" is actually the best of the options mentioned so far in this discussion. "Eligibility" is way too vague (neither an aim nor a criterion) and "citeability" is just wrong (that would refer to sources, not topics). The word that has annoyed me the most, for the past 20+ years, is "verifiability", which in wikispeak means something entirely different from its meaning in plain English. In plain English, something is verifiable if its truth can be confirmed, which is why the ancient slogan "verifiability, not truth" is my nomination for the worst own-goal in Wikipedia history. Zero^talk 10:35, 21 February 2026 (UTC)

  In defense of verifiability, it's purely a question of which frame of reference you're using. Sure, in everyday speech it usually means verifiable [against objective truth] but it's not a stretch or corruption of the meaning that on Wikipedia it means verifiable [against a source text]. The whole point of 'verifiability, not truth' is to clarify that the thing we're verifying against is not objective truth, merely source->text integrity. Athanelar (talk) 10:48, 21 February 2026 (UTC)

  No, that was not where "verifiability, not truth" came from (I was here when it was adopted). The "not truth" part refers to "no original research". The idea is that we use what reliable sources say is true and not what we personally believe is true. It isn't a reference to objective truth. The problem with the slogan is that it was commonly taken to mean that Wikipedia doesn't care about getting the facts right, and this misunderstanding got "out there" to our detriment. And we threw it at newcomers before they had a chance to grasp that "verifiability" didn't mean what they thought it meant. Zero^talk 11:40, 21 February 2026 (UTC)

  When it comes to "verifiability" I view it as "is the notability verifiable." Something can be true, but not notable. There is a lot of stuff about me floating on the internet, my existence is verifiable, however none of it meets the criteria for notability, my notability is not verifiable. GeogSage ^{(⚔Chat?⚔)} 20:01, 23 February 2026 (UTC)

I would probably call it "sourceability" which is somewhat more accurate. However, as said before it's one of these entrenched terms that are hard to change. Jo-Jo Eumerus (talk) 14:25, 21 February 2026 (UTC)

What name is being proposed, to change "Notability"? GoodDay (talk) 14:37, 21 February 2026 (UTC)

Nothing currently; I'm not trying to make a proposal, just to discuss the topic. There's no point in proposing a candidate if nobody thinks it should be changed to begin with. Athanelar (talk) 14:42, 21 February 2026 (UTC)

No-one can decide. And most likely no decision will be made. This is such an obvious waste of time I don’t really know why I, or any of the many high-profile editors here, dignifying it with a response beyond “WP:PERENNIAL” Dronebogus (talk) 14:43, 21 February 2026 (UTC)

I note WP:PERENNIAL doesn't actually have this topic listed (yet). Anomie⚔ 20:57, 21 February 2026 (UTC)

"Notability" is fine. It means what subject can be noted on Wikipedia. I don't see a glaring problem with it. Joe vom Titan (talk) 14:46, 21 February 2026 (UTC)

The glaring problem is that that's an extremely unintuitive and niche use of that word, which usually means "important" or "significant" Athanelar (talk) 16:34, 21 February 2026 (UTC)

That's what "notable" means, while "notability" reflects how much someone is likely to be notable, which matches the intent of what WP's notability does. Masem (t) 16:47, 21 February 2026 (UTC)

But "importance" is subjective (my wife is important to me, the road I take to work is important to me, the band my friend started when he was 15 is important to him; none of them are notable in the Wikipedia sense). "Notability" (or at least the GNG) aims to be an objective standard that is either met or not and has little to do with what most people think of as "importance". HJ Mitchell | Penny for your thoughts? 20:49, 21 February 2026 (UTC)

Also worth adding that we do "note" unnotable things on Wikipedia, just within articles rather than as standalone topics. CMD (talk) 04:18, 22 February 2026 (UTC)

Its far from an objective standard, which is why notability is a rebuttable presumption. Show that the topic is given in-depth coverage from at least a few independent, reliable sources (generally being secondary sources), and we'll presume that the topic can merit a full article. But there's so much variability in what qualifies as in-depth coverage, how many and what kind of sources, etc. that its far to call the test solely objective. Otherwise, we'd not have any problem at AFD with deletion.

But we do associate being notable as if the topic was important enough to independent authors to cover in-depth, that is, is the topic demonstrated the quality of being notable based on sourcing. Masem (t) 04:29, 22 February 2026 (UTC)

When you read that notability is necessary for inclusion on Wikipedia how is it not glaringly obvious that it is Wikipedia that sets the standards for what is notable?--User:Khajidha (talk) (contributions) 18:09, 22 February 2026 (UTC)

While we’re here, why don’t we look at all the other less-than-ideal names used for rules on Wikipedia? WP:NPOV (which isn’t neutral) WP:IAR (don’t actually do this) WP:DELETION (pages aren’t deleted). I could probably find lots of examples. Wikipedia is just like any hobbyist subculture in that it has a lot of weird jargon that doesn’t necessarily mean what the dictionary and common sense say it means. “Fixing” that will just create more problems as now both newbies AND veteran editors are confused by the weird new terminology. On top of that Newbies still won’t understand what it’s meant to convey, veterans will just keep using the same terminology they always used, and eventually it will just get reverted back with the same unnecessary cost as changing it. Dronebogus (talk) 15:00, 21 February 2026 (UTC)

• Notability is a terrible name because it's easily conflated with "importance", which is subjective—everything is important to someone. I've previously advocated for "criteria for inclusion". HJ Mitchell | Penny for your thoughts? 15:20, 21 February 2026 (UTC)

  One word synonym could just be "Includable." GeogSage ^{(⚔Chat?⚔)} 20:03, 23 February 2026 (UTC)

  It looks like we have a nice set of redirects in the Wikipedia:Criteria for inclusion line. I think that's a good thing. People can use whichever they like. WhatamIdoing (talk) 23:04, 28 February 2026 (UTC)

• I prefer "Notability". While "criteria for inclusion" would convey the idea, it would be awkward to use regularly. None of the other suggestions above work for me. - Donald Albury 19:04, 21 February 2026 (UTC)

  My POV hasn't changed since the discussion last year, specifically that we should eventually change this, and that the way to go about it is to pick some other word or phrase and use both, e.g., "On Wikipedia, notability, or eligibility, is a test used by editors..." or "On Wikipedia, notability is the article creation criteria that editors use..." Then editors have a choice, and if they choose to say "It's Wikipedia:Notable" or if they choose to say "It meets the Wikipedia:Article creation criteria" or if they choose to say "I think this meets our Wikipedia:Eligibility standards", then that's fine (though it'd be preferable if the guideline suggested a single alternate name). WhatamIdoing (talk) 23:02, 28 February 2026 (UTC)

Hey @Athanelar, I'm the person who started the last massive discussion. Good luck. The main message I came out of it with is there will be many more opposing people in actual RfCs rather than discussions; I started an RfC thinking I would have significantly more support than I did based on my experience discussing it at the idea lab. I think the only way to make this work is to make a smaller change first -- maybe some sort of movement among AfD contributors to use eligibility (linking to notability) would work to get it off the ground, but I have no idea how that would be organized. Maybe a WikiProject? Mrfoogles (talk) 20:59, 27 February 2026 (UTC)

• Can we at least footnote the WP:N lead sentence with an explanation along the lines of:

  (1) Wikipedia notability is largely independent of real-world notability, (2) while this is confusing, we continue to use the word because multiple discussions have failed to find a better one, and (3) alternative names (that have been considered, but not adopted) include notedness, criteria for inclusion, eligibility, suitability, admissibility, and wikinotability.

  New editors start out assuming that Wikipedia notability is at least somewhat related to real-world notability, which isn't helped by WP:N statements like Determining notability does not necessarily depend on things such as fame, importance, or popularity. While new editors don't start by reading all the PAGs, WP:N and WP:GNG are quoted so often they'll likely see them first, making it even more important these pages clarify common misconceptions. An overview of previous discussions will also be of value to more experienced editors. Preimage (talk) 06:41, 1 March 2026 (UTC)

  Alternatively, we could remove the term altogether (in effect, making WP:N a self-referential acronym): "On Wikipedia, WP:N is a test used by editors..." Preimage (talk) 07:08, 1 March 2026 (UTC)

  Please note that the existence of the term "Notability" is essential to a joke on the signpost. I think it was in the comix section of the last January edition. ~2026-11404-95 (talk) 19:51, 3 March 2026 (UTC)

• I dunno why people have complained how terrible "notability" is other than... probably it's unfair to those who may not be "notable" but might deserve an article perhaps. This is more akin to (failed?) efforts to repeal and (failed) constitutional challenges to the Affordable Care Act, both perhaps time- and money-wasting. Right? Frankly, "notability" has been fine as-is, despite hostile backlash and all, and something that consensus should practice often. Too bad certain others here wanna change it. BTW, have standards of "notability" been that low or that high? George Ho (talk) 19:55, 4 March 2026 (UTC)
• I like Preimage's solution of referring to Notability as a native concept abstracted from outside usage, but I like better the idea of changing the word notability. I've created WP:WOTABILITY to try to best differentiate outside notability with WP's notability. If anyone has any better idea than my sort of clunky one please share - I suspect this may be a big problem in editor retention, to have such an onerous stumbling block placed so early in editor lifetime. Embyarby (talk) 04:02, 5 March 2026 (UTC)
• I am not sure what is wrong with the word notability. There are lots of citable subjects that are not really notable. Citability (is that a word) does not necessarily mean that the subject alone is notable, unless of course those citations come from notable or established sources. Wikipedia has much bigger issues right now than the usage of notability for establishing subjects. Words in the Wind(talk) 21:11, 9 March 2026 (UTC)

While I very much agree with this, I'm not sure what singular/compound word could replace it.

At AfC, I've started to refer to Notability to newcomers as "what Wikipedia calls 'notability'", or just saying "hasn't been covered in multiple reliable, secondary, independent sources". Nothing will scare someone off for good than saying that the topic they're writing about isn't notable, or perhaps even that THEY themselves aren't notable, in the case of autobiographies (which can very easily be taken as a passive-aggressive insult!).

I do agree with @Preimage - putting something at the top of WP:N to differentiate between real-world notability would be good. Or maybe even a change to the "This page in a nutshell" banner.

I do think a newbie friendly page to the notability guidelines could work out. It would be more detailed than Help:Introduction, but less jargony than other P&G pages. Pretty much, a line-by-line breakdown of key points like WP:GNG, WP:NTEMP, WP:WHYN. EatingCarBatteries _{(contribs | talk)} 05:06, 10 March 2026 (UTC)

"Notability" is clearly the best option. As Masem noted, "eligibility" is awful because it implies that notability is binary, which is only true if you treat the blurry line between inclusion and exclusion like an obnoxious WP:WIKILAWYER. "Notability" clearly means notable to Wikipedia – just like events notable to Kotaku and to The Law Society Gazette will have minimal overlap. Theoretically, that concept is capturable in "Relevance", which I haven't seen discussed here yet, but (just preempting this, because the anti-"notability" camp is ostensibly desperate for any alternative) this is even worse: 1) it's a lateral move at best because we're a general-purpose encyclopedia, and 2) it would completely overload the common word "relevant" across Wikipedia.

Having read this entire discussion, the suggestion is well-intentioned but nonsensical bikeshedding. No better term has been put forward (because, in my opinion, it can't be – unless we all decide on "cromulent" and use our hivemind to collectively understand it), the concept has already been baked-in for over 20 years, multiple attempts to change it in the past have ended in failure, the definition is literally right there and plastered around any discussion thereof if there's any confusion, the consequence of misunderstanding it is excruciatingly low-stakes, most misunderstandings of any real consequence come from not reading guidelines that a word or three could never capture on their own or meaningfully encourage someone to read, and in 99.9% of cases, it comports with the lay meaning just fine anyway. TheTechnician27 (Talk page) 01:28, 13 March 2026 (UTC)

Notability/Eligibility/whatever you want to call it should sound binary, because whether we have a Wikipedia:Separate, stand-alone article for a given subject is also binary.

(Also, did you read the literal "definition"? The one that says "Notability is a test"? Notability is not a test.) WhatamIdoing (talk) 05:17, 13 March 2026 (UTC)

I prefer to treat the issue of notability as; "is notable::notability has not yet been established". There will always be a grey area, border notability, where we will argue over the notability of a topic, a point where finding one more piece of significant coverage in a reliable source might push the topic over the line to notability, or an obscure topic captures enough attention in the real world to result in new significant coverage in reliable sources (I repeatedly reverted attempts of a certain musician to add themselves to Wikipedia until one day I saw that they had finally made enough of a splash to get significant coverage in reliable sources). Donald Albury 16:53, 13 March 2026 (UTC)

There are 2 pages that in conflict: TLC and NSYNC. Both have pre-debut members who never released any song as the group member. The problem is, some Wikipedia editors list them as past members. In this case, Blackpink's page should also include Miyeon, BTS with Supreme Boi, Big Bang with Jang Hyunseung, SNSD with Soyeon, and Innosense with Britney Spears.

Now, what is the consensus? ~2026-41110-5 (talk) 14:18, 23 February 2026 (UTC)

I don't see the problem. As long as the group has maintained the same name, then those pre-fame members are still former members of the group. There should definitely be a clear delineation of when a person was a member of the group in the prose and there may be special organization or footnotes in lists or tables, but they still need to be in there. --User:Khajidha (talk) (contributions) 17:28, 23 February 2026 (UTC)

Reading the further responses, I see some possible distinctions to draw. A group of musicians may come together and practice. At some point, they try out for bookings and print materials to advertise themselves. Eventually, they begin booking shows. They finally receive a recording contract and release their debut album. People may come and go at any point in this sequence. My position is that anyone who is presented to the public (either in marketing materials or actual performances) should be counted as a member. Where in this sequence do the people you mentioned fall? --User:Khajidha (talk) (contributions) 13:35, 25 February 2026 (UTC)

I don't know exactly why the pages for TLC and NSYNC have the very early members listed, but in the case of the K-pop groups listed; just absolutely no. You'd need to add over half a dozen "members" just for SNSD alone, because leading up to their debut, their management was swapping potential members in and out at seemingly a moments notice. There were at least four different line-ups prior to debut, with as many as twelve members in them. And that's just the ones we know of, I'm sure there were plenty more, are we gonna add every SME trainee from that period as a potential past member? Which brings me to a further point: good luck finding reliable sources any of this, because every thing I could find in my research for this reply was terrible fansites, or "sources" known to be unreliable. Pre-debut trainees are just that: trainees. They are not official members.DragonFury (talk) 22:28, 23 February 2026 (UTC)

I mean, the actual final plan of SNSD was 10-member group before Soyeon left voluntarily.

But yes, pre-debut members should not be included since it's not official. I also don't know why they including pre-debut members on NSYNC and TLC. ~2026-41110-5 (talk) 22:34, 23 February 2026 (UTC)

And to add another point; Twice was formed from the TV program Sixteen, so do we add the seven people not selected in the program as past members? Because if so, I'd like to see someone make the same argument for Kep1er and the NINETY contestants from Girls Planet 999 who weren't selected for the group. DragonFury (talk) 22:38, 23 February 2026 (UTC)

With TLC, it's because 1) she's mentioned in reliable sources as 2) the founder of the group. Similarly with NSYNC, it's because 1) Jason is mentioned in reliable sources as 2) a formative member of the group. We should default to sources here. Is someone reliably claimed to have been a member of the group? Then they should be included. Now, should they always be in the infobox? Should they never be in the infobox? No. That's editorial judgement, and it depends on the band. And I don't think a recording contract or "official" debut release is a good enough rule of thumb. Someone could be an important member of the group without that. Some groups release music without a contract. And judging "importance" here is often arbitrary, anyway.--3family6 (Talk to me|See what I have done) 21:03, 14 March 2026 (UTC)

It should be whatever is reflected in reliable sources. If reliable sources describe those people as actual, bonafide former members of the band then sure.

For most of the examples you listed, however, this wouldn't be the case... they were just trainees in the same company along with the people who eventually became the final lineup. Agencies add and remove people from lineups all the time before they actually release any music. It doesn't mean they're a member of the band. RachelTensions (talk) 01:41, 24 February 2026 (UTC)

so you agree that we should not include pre-debut members? ~2026-41110-5 (talk) 19:44, 24 February 2026 (UTC)

I'm going to quickly reject "never released any songs as a member" as the dividing line, because bands can have a healthy life as a performing group before they become a recording group. There were no The Beatles songs released with Stu Sutcliffe on them until he was dead over thirty years (if I recall correctly), but he was a key player in their pre-record-contract days. -- Nat Gertler (talk) 02:16, 24 February 2026 (UTC)

The Beatles does not include Sutcliffe, Pete Best, etc in the infobox -- instead they add a link to the former members section, so that would be a solution for this kind of issue. ^~Hydronium^~Hydroxide^~(Talk)~ 04:18, 24 February 2026 (UTC)

Based on your contribution, this discussion seems to originate from a content dispute involving a temporary account that has since been blocked. As such, it is unclear why the argument is being revisited here rather than being handled on the relevant article's talk page, or whether it is intended to validate that blocked account's viewpoint. Regardless, there is general mutual understanding among most editors that only individuals who officially debuted and/or released work with the act are listed as members. Those involved solely during pre-debut stages, such as auditions or trainings, are typically excluded and do not carry long-term encyclopedic significance in relation to the act. Such information, if sourced reliably, may instead be included in the individual's biographies instead, should they have an article here. This standard has been applied consistently across comparable articles, with limited exceptions that are likely the result of per-article consensus. — 🧧🍊 Paper9oll 🍊🧧 (🔔 • 📝) 03:16, 24 February 2026 (UTC)

So the consensus is only official debuted members are listed, right? ~2026-41110-5 (talk) 07:33, 24 February 2026 (UTC)

For clarity, when I refer to "general mutual understanding", I'm specifically referring to South Korean musical acts only. That wording is intentional and is applied in conjunction with WP:VERIFY. As for anything else, refer to the last sentence in my earlier reply above instead. Additionally, it appears you are conflating two different categories of group formation. Certain Western acts where membership may be tied to founding history, name origin, or early lineup evolution are being treated as directly comparable to K-pop acts, where membership is defined through a formalized debut system and post-debut activity. Applying the same framework to both assumes a one-size-fits-all standard, which does not reflect how these industries function. Because of that conflation, comparisons such as invoking NSYNC to justify listing pre-debut trainees in K-pop groups are not particularly meaningful. For cases outside the K-pop context, exceptions and differing treatments may and or could already be accounted for on a per-article basis, or may not be, as noted previously. — 🧧🍊 Paper9oll 🍊🧧 (🔔 • 📝) 08:25, 24 February 2026 (UTC)

I mean, “pre-debut” means before debut, right?

Before debut = unofficial. Does it mean the same everywhere, or does each country have its own definition of “pre-”? Doesn't matter. It’s still unofficial.

I feel like Wikipedia is turning into a Wiki Fandom now. Why are we including unofficial members / non-final members in the same list as the official members? Where’s the standard? What makes Wikipedia different from Wiki Fandom? It's unofficial, like how I explain it better to you all. It's unofficial. ~2026-41110-5 (talk) 11:07, 24 February 2026 (UTC)

Alright let's summon User:Binksternet who putting Jason Galasso as former member of NSYNC. In his defense, Jason's name was part of the NSYNC name (which is already changed to Lansen). ~2026-41110-5 (talk) 07:47, 24 February 2026 (UTC)

• Why are we having this content dispute on VPP?—S Marshall T/C 10:23, 24 February 2026 (UTC)

  Because it's still about guideline. We are literally talking about an undisclosed guideline for hundred pages related to musical groups. ~2026-41110-5 (talk) 11:27, 24 February 2026 (UTC)

  No we aren't. This decision is fact-sensitive and needs to be made on a case by case basis. It depends on the person's contribution -- did they help write a key song, did they affect the band's development or sound, were they just a hired gun for a few shows? What do the sources say about them? There's no guideline to write. It's a simple matter of editorial judgement. If there's dispute then they don't go in the infobox (WP:ONUS).—S Marshall T/C 12:28, 24 February 2026 (UTC)

  Exactly that's why we are here. You think we are here for planning slumber party? No, we are here for discussion. And there is a big disagreement about whether people should include pre-debut members or not. Did those pre-debut members help the group? Put them on the History section, not the main infobox. That's the function of History section, to tell the story behind the group's creation. Infobox is only for the official information. ~2026-41110-5 (talk) 19:42, 24 February 2026 (UTC)

  Official in whose eyes? Is there an office somewhere where a team of people decide who is official or not? Without answering that queston it's no good bolding "it's unofficial". There are so many differences between times, places, genres and individual bands that it's pointless trying to create a general rule that will apply to everyone. And it seems that by pre-debut you mean before the recording debut. Many notable bands don't make recordings until a few years into their careers. Phil Bridger (talk) 19:42, 25 February 2026 (UTC)

  "official in whose eyes" oh my goodness. Wikipedia is using what kind of source for reference? Reliable sources, right. So it should be according to reliable sources. What those reliable sources say about "official members of the group"? ~2026-41110-5 (talk) 03:09, 9 March 2026 (UTC)

  I would be surprised if there were reliable sources on American bands that made a distinction between "official" and "unofficial" in the way you describe, and I'd expect them to simply describe the people you're trying to remove as having been members of the band. Sesquilinear (talk) 03:23, 14 March 2026 (UTC)

  Let's change the definition from "recording debut" to the time they sign the contract. That's the exact situation when the member is the official member of the group. ~2026-41110-5 (talk) 03:13, 9 March 2026 (UTC)

The sources talking about the formation of NSYNC all agree that the name of the band came at a time when bass singer Jason Galasso was in the group. Galasso was warmly received when he joined, with other group members telling him his voice was what they were finally looking for. It was this formation that was deemed ready for a record label deal. Timberlake's mother Lynn remarked that, with Galasso on bass, the group now sounded very much "in sync". They noticed that the final letter of each group member's name could be assembled to form the name NSYNC, with Jason supplying one of the letters. That means Galasso was a foundational member of the group. Supporting sources include a Timberlake bio book and a People magazine piece on Galasso, calling him the "original fifth member of NSYNC".

I don't think there is any basis for artificially drawing a line for the purpose of rejecting Galasso. We should not be determining band membership on our own by saying "pre-debut" or whatever. That would be a violation of WP:No original research. The reliable sources should be telling us who are the group members. Binksternet (talk) 20:40, 2 March 2026 (UTC)

I mean, he was the original member YES there are reliable sources telling that he was the original member. But that's not the point of this discussion.

Looking at the article, it literally says that he backed out the plan to debut, so he didn't sign the contract. After he left, the second N word for NSYNC is placed to Lansen (Lance Bass). So basically there's no Jason in the name, his name is replaced by Lansen.

The point of this discussion is about whether we should include pre-debut members in the main infobox, as pre-debut members are unofficial members. Why putting people that didn't sign the contract to the main infobox? They are unofficial. Unofficial. Should i repeat it for the third time? ~2026-41110-5 (talk) 03:07, 9 March 2026 (UTC)

Bands and groups exist before signing contracts and after losing contracts. --User:Khajidha (talk) (contributions) 13:13, 9 March 2026 (UTC)

You're creating an arbitrary, or at best industry-specific, definition of "official" and "unofficial" as a blanket rule for every band page. As has been said repeatedly, people can be members of bands without a formal contract or releasing a recording. We go by what sources say.--3family6 (Talk to me|See what I have done) 20:53, 14 March 2026 (UTC)

I think the problem with this is that the OP seems to think that one tiny genre of popular music, where terms like "unofficial" and "pre-debut" have some meaning, is representative of the whole. Phil Bridger (talk) 21:21, 14 March 2026 (UTC)

It seems a third of the sources used to train AI is material from Wikipedia. Shouldn't something be done about that? We all are doing this for free to inform people, not to support billionaires AI investments who can push out the people of their daily jobs. I've been working on several Wiki's in different languages, and I have the feeling my work is stolen by big companies. — Preceding unsigned comment added by ~2026-13646-72 (talk) 19:49, 2 March 2026 (UTC)

1. Do you have sources to prove that "a third of the sources used to train AI is material from Wikipedia)?
2. Who cares if we're "supporting" billionaires? Not only is that a major oversimplification of the AI debate (more people use and are affected by AI than just the billionaires), but you also support billionaires through the brands you use in your day-to-day life (such as the device that you use to edit Wikipedia on, which was made by a company run by a billionaire), so the billionaire point doesn't hold much weight.
3. Where is the evidence that AI hasn't created more jobs than have lost, and what about the people whose jobs have been improved by AI (such as programmers, engineers, or medical researchers)? Besides, ending jobs doesn't make something automatically wholly evil; the internet has put a lot of jobs at stake (online shopping, for example, has put many brick-and-mortar stores on the line), yet that does not mean we should just forget about the benefits of the internet. The same thing can be said about AI.
4. How is AI analyzing information to train its ability to make something completely different "stealing?" Besides, your work on Wikipedia isn't wholly "your" work; not only have many other editors contributed to the articles that you have wrote on, but you and other editors have gotten "your" info from other sources, so it's not like AI is "stealing" your personal, wholly-original creation. Besides, Wikipedia is not meant to be anyone's source and its use is not meant to be restricted; it's meant to be a free-to-use platform for anyone to gain information. Why does that stop with AI analyzing that information? Furthermore, why doesn't it stop with others using Wikipedia as a source to get information for their own projects? I don't think you would get mad at someone on YouTube "stealing" from Wikipedia articles to get the info they need for their video essay.

QuisEstJoe (talk) 21:53, 2 March 2026 (UTC)

When you edit Wikipedia, you are agreeing to allow your work to be used by anyone in any way, basically. We don't have a way of excluding big companies (and indeed, some huge ones like Google use it in very obvious ways), and it would not serve the stated goals of the project if we did start excluding. -- Nat Gertler (talk) 22:00, 2 March 2026 (UTC)

not "in any way"

they still have to abide by either Wikipedia:Text of the Creative Commons Attribution-ShareAlike 4.0 International License or Wikipedia:Text of the GNU Free Documentation License

and as far as I know, they would have to add some notice to every LLM output since some fraction of it is a derivative work of content released under those two licenses, which mandate that you state who the original author(s) was/were Laura240406 (talk) 00:32, 10 March 2026 (UTC)

What could be done is calls for and work on better attribution/linking to Wikipedia as said earlier. However, whether it's really required isn't so simple and probably it wouldn't be unless it's a quite direct quote. People can read Wikipedia and other sources and then describe sth in more or less their own words without having to attribute all those sources. LLMs can do the same: usually it's not direct quotes or translations. It's true that "in any way" is false though and I would prefer if whenever a Wikipedia article is one of the main knowledge sources of an LLM output, it would link to the respective article(s). Prototyperspective (talk) 13:23, 10 March 2026 (UTC)

The C-by-C license covers the use of copyrighted material. Facts and ideas cannot be copyrighted, only the expression of those facts and ideas can be copyrighted. Anybody can reuse facts and ideas however they want, as long as they do not copy the expression of those facts and ideas. As long as they are not copying the expression of facts and ideas extracted from Wikipedia, the requirements of attribution under the C-by-C do not apply. We the editors own the copyright on content we add to Wikipedia, but we do not own any copyright on the facts and ideas in that content. Donald Albury 17:01, 10 March 2026 (UTC)

I am not a fan of AI, but one of the pillars of Wikipedia is that all editors freely license their work to the public, and no editor owns an article – any contributions can and may be mercilessly edited and redistributed.

Whatever wikis you're working on probably have AI cleanup efforts you may be interested in. Gnomingstuff (talk) 02:47, 3 March 2026 (UTC)

I am not agreeing to allow my work to be used by anyone in any way. Wikipedia:Reusing Wikipedia content: Wikipedia's text content ... can be used under the terms of the Creative Commons Attribution Share-Alike license (CC BY-SA)... To re-distribute a text page in any form, provide credit to the authors... If you make modifications or additions to the page you re-use, you must license them under the Creative Commons Attribution-Share-Alike License 4.0 or later. Outside of that, the copyright reverts back to me. Hawkeye7 (discuss) 03:22, 3 March 2026 (UTC)

I desire to see a source for your claim. Also, you agree by using Wikipedia that "agree to irrevocably release your text under the CC BY-SA License and GFDL." And "You agree that a hyperlink or URL is sufficient attribution under the Creative Commons license." This means that an AI agent providing a link to the Wikipedia page is enough to disqualify you from saying your work was stolen. You did agree to allow your work to be used. ~2026-11404-95 (talk) 19:48, 3 March 2026 (UTC)

This is what it means to license your work under a copyleft licence: you can't restrict who reuses your work if they meet the terms of the licence. And something is being done, by Wikimedia Enterprise getting Big Tech to pay for high-volume access without taking down the servers. See the various news reports following their announcement on our 25th birthday, such as this one in Ars Technica. ClaudineChionh (she/her · talk · email · global) 21:45, 3 March 2026 (UTC)

Hi ~2026-13646-72

The "one third" number needs a source, but yes, AI companies scrape Wikipedia for training data, this is a known fact. They are allowed to, under the CC BY-SA license, but other editors have identified that there is still the issue of attribution. As I understand it (I am not a lawyer), the considerations for licensed works used for training data is a legal grey area. And I agree, it does suck, because I agree that the AI industry sucks.

There's two senses in which you could say Wikipedia could "do something" about this. You either mean the Wikipedia community (the volunteer editors who edit Wikipedia), or the Wikimedia Foundation (the non-profit that hosts the website itself). I don't have much to say for what the Wikimedia Foundation could do.

For what the community could do, Wikipedia isn't really here to right great wrongs; we don't organize protests or anything, we're just here to build an encyclopedia, and we certainly shouldn't artificially inject anti-AI bias into the encyclopedia. You can, however, help chip in over at the AI cleanup efforts! We do (on the whole) consider using LLMs to contribute to Wikipedia to be abusive, but many folks try it anyways, and it takes no small amount of volunteer effort to deal with that abuse. You could always become an editor and assist over at WP:AINB. MEN KISSING (she/they) T - C - Email me! 20:10, 4 March 2026 (UTC)

I honestly don't even think that it sucks that much because Wikipedia is supposed to be a free-to-use encyclopedia for everyone to use, and suddenly closing that off to AI training itself because it's using "stolen" data is both unfair and nonsensical. If you come to Wikipedia to create "your own" article for yourself (which, just to be clear, I'm not saying that you in particular are), you're kind of missing the point of Wikipedia. Besides, it'd be one thing if this was an art-sharing or fiction-writing forum where we publish artistic expression, but saying to someone training AI that they can't use an article of real-world facts because "that's MY article of facts that I have the right to" is simply selfish. QuisEstJoe (talk) 20:40, 4 March 2026 (UTC)

Well, the sentiment against AI scraping of Wikipedia is not only based on the idea of content ownership being violated. It's a complicated issue with a lot more moving parts than that.

This more gets into the wider politics surrounding AI, and I'm sure that can be quite polarizing, so I'm not too eager to dig into it. But perhaps that wider conversation is one we should have, because some of the involved politics is quite relevant to the Wikipedia project. MEN KISSING (she/they) T - C - Email me! 22:32, 4 March 2026 (UTC)

We pause to remember the Great Wikipedia Black Out of 2012. Hawkeye7 (discuss) 00:40, 10 March 2026 (UTC)

As others have explained above, whether you're fond of LLMs or not, it would be antithetical to the purpose of Wikipedia for us to impose restrictions on who can use its content, and for what purpose. That is the entire point of calling Wikipedia the "free encyclopedia"—not only in the sense that anyone can edit it, but also in the sense that anyone can use it, reuse it, reproduce it, redistribute it, etc. Kurtis ^(talk) 14:48, 5 March 2026 (UTC)

Shouldn't something be done about that? What could be done is calls for and work on better attribution/linking to Wikipedia. We all are doing this for free to inform people, not to support billionaires AI investments who can push out the people of their daily jobs. […] I have the feeling my work is stolen by big companies. Speak for yourself.

1. automating work is great success and gets us closer toward a society where we have enough people working on meaningful jobs and in areas where workers are missing such as renewable energy industry and care work
2. it's not supporting billionaires but at least also lots of people who benefit from these technologies; additionally there's free open source AI software that's not supporting billionaires – nuance please, it depends on which AIs you use how
3. I think it's amazing if my volunteer work becomes more useful and reaches larger audiences, so instead of feeling 'my work is stolen' I feel like my work in Wikimedia projects is more worth my time and effort.

Prototyperspective (talk) 17:40, 9 March 2026 (UTC)

Say, for example, you have an editor who's written an article about XYZ. It is well-written, well-sourced, comprehensive, and it's about an actual person who actually lived. They get it to GA-status, and eventually, to FA-status. But later on, they admit that they used ChatGPT in helping them to create that article. They explained the process that they used in doing so:

• First, they found a bunch of reliable third-party sources about the subject in question.
• They use the following prompt to get ChatGPT to generate text: "Create a comprehensive overview of XYZ using the copy-pasted text provided below the horizontal line. Format this overview in the style of a Wikipedia article."
• They double-space, add three underscores (the "horizontal line"), and then copy large chunks of text from the various sources they've assembled, sectioning them according to the source from which they were excerpted.
• ChatGPT generates the text of a Wikipedia article; the user copies and pastes it into Google Docs.
• They then begin using their sandbox and the preview button to prune, paraphrase, copyedit, fact-check, expand, and further "Wikify" what ChatGPT generated.

By the time they're through with their work on the article and are ready to submit it, it can no longer be readily identified as AI-generated—it is almost entirely in their own words, using sources that they compiled. They checked it for factual accuracy, and they made sure that everything flowed together smoothly.

In your view, is this still a problematic use of generative AI for the purpose of editing Wikipedia? If so, why do you consider it problematic? And if an editor confessed to doing something like this, do you feel they should face sanctions for having done so? Kurtis ^(talk) 15:23, 5 March 2026 (UTC)

• Your example isn't realistic. If a person is going to rewrite an LLM-generated article so that every word in it is their own, then what's the purpose of LLM generation in the first place? ~2026-14165-49 (talk) 17:08, 5 March 2026 (UTC)

  Exactly. See my post below. EEng 17:16, 5 March 2026 (UTC)

  They would use it as a springboard to actually building a proper article, and to make the whole task feel significantly less daunting. Kurtis ^(talk) 17:17, 5 March 2026 (UTC)

  It can only "feel significantly less daunting" if they skip all the hard parts of writing an article i.e. gathering and internalizing the sources, selecting the facts for presentation, deciding on article organization, and so on. Your "springboard" is, in fact, a nice way of saying, "A way of letting people who aren't able to usefully contribute, feel like they're usefully contributing". Things like you're proposing will be the death of Wikipedia. EEng 17:32, 5 March 2026 (UTC)

• It's problematic because it's fiction. The fact-checking part would never happen, because by the time you prune, paraphrase, copyedit, fact-check, expand, and further "Wikify" what ChatGPT generated you could have just written the article from scratch with less effort (not to mention less chance of AI excrement slipping through). Anyone with sufficient competency to do what you describe would know this, and so would know better than to actually do it; therefore those who claim they're doing it actually aren't competent to do so (even if they fool themselves into thinking they are, in fact, doing so -- Dunning-Kruger). That's why there's no use case for AI-generated content, and why those who use AI to generate either article content or talk-page posts are, ipso facto, incompetent and should be blocked on sight. EEng 17:16, 5 March 2026 (UTC)

The reason I would think of for someone doing this, even if it ultimately expends the same amount of effort as simply writing the article from scratch (if not more), is because it might make the act of actually writing an article soup-to-nuts feel less daunting. Logically, anyone would surmise that they're just doing content creation in a roundabout way, but the sight of a "completed" article with flaws might give them more of an impetus to develop and build upon what was generated. It could create the illusion of editing rather than building an article from the ground up. Kurtis ^(talk) 17:27, 5 March 2026 (UTC)

"Illusion of editing"? You want us to foster an "illusion of editing"??? EEng 17:33, 5 March 2026 (UTC)

Was it not clear what I was attempting to convey? By "illusion of editing", I meant that it would feel more like editing an article as opposed to writing an article from scratch, even if that is functionally what it is. Kurtis ^(talk) 17:41, 5 March 2026 (UTC)

No, it was not clear. In general, most "editing" (in the sense you're using the word) involves improving presentation, copyediting, maybe pruning out UNDUE cruft, by making incremental changes to an article which is assumed, by default, to be verifiable, free of copyvios, and otherwise in conformance with key policies. What you're proposing is that we allow an inexperienced user be presented with a ton of material which without question is studded with hidden crap, and expect that user to find and remove all that hidden crap. EEng 23:06, 5 March 2026 (UTC)

Oh, I disagree with this. Sometimes it's useful to have a sort of template to start from when writing an article. The AI can basically work like an outline as you do your own research. This is fine as long as the end result is in your own words and accurate. Sentimental Dork (talk) 19:27, 9 March 2026 (UTC)

Say, for example,

Nope, start over and try again. Which specific Featured Article are you vaguely alluding to in this hypothetical?

Anyway, your thought experiment here has a lot of problems:

• WP:LLMDISCLOSE strongly recommends that people disclose AI use from the get-go, in all associated edit summaries. This is not an outrageous, time-consuming thing to do. Here's an example of someone doing it usefully in good faith.
• Your sense of how the average AI Featured Article comes about is unrealistic. Based on the cases I've seen, the more common scenario is that AI is used to find the sources (potentially introducing non-reliable sources/WP:UNDUE emphasis/bias), to summarize those sources (likely introducing source-to-text integrity issues), and copyediting/rewriting existing text (potentially introducing tone issues, and distortion of meaning).
• There's no way to prove the negative here, but many articles that people think "can no longer be readily identified as AI-generated" can be identified as AI-generated in about 15 seconds if you know what you're looking for. I've flagged several GAs and FAs as AI-generated (and was right), and there are several more GAs and FAs that I'm pretty sure contain AI text but haven't bothered to flag because I don't really feel like getting into even more arguments than I already do. (Panic Room, it's your time to shine!)
• There's no way to prove how much fact-checking people do, but in my experience, when I spot-check sentences and claims that obviously came out of an LLM, the citations do not often verify the text, and sometimes they don't even come close. WikiEdu found last year that of the AI-generated articles they flagged, nearly every cited sentence failed verification. Similarly, a lot of AI text has a sheen of promotional tone that, for whatever reason, just doesn't register on some people's radar no matter how obvious.

Gnomingstuff (talk) 18:10, 5 March 2026 (UTC)

@Gnomingstuff: Which specific Featured Article are you vaguely alluding to in this hypothetical?—It really is just a hypothetical, but in the interests of disclosure, it's an idea that I'd been mulling over in recent times. Over the past few years, my executive dysfunction due to ADHD-PI has worsened (I suspect due to Long COVID-related brain fog), and so I have difficulty motivating myself to write much of anything. I've played with the idea of using ChatGPT to create a sort of "skeleton" for an article, or for generating something that I could use as inspiration. I want to be very clear: if I add something to Wikipedia, you can rest assured that it will have been thoroughly fact-checked and well-sourced, and each claim that has a reference attached to it is directly and explicitly verified by that reference. However, I was unsure of how the community would respond if I were to write an article, and then disclose that I had used ChatGPT to lay down a "blueprint" of sorts. I genuinely think that for a lot of editors, this would be seen as a serious breach of trust, even if what I submit is qualitatively different from what ChatGPT spat out; it effectively circumvents much of the article-writing process. I'm not sure if an editor who did this would be blocked for doing so, but I do think it would cast their integrity and their critical thinking skills into doubt, and it might trigger a review of whatever content they'd submitted using this approach.

Your linking of LLMDISCLOSE demonstrates that current community practice is more nuanced than I would have expected. Disclosure of using an LLM strikes me as the ethical thing to do, but it's fraught with risk. A lot of editors who attempt something like what I've described would probably opt against doing so, as it would undermine trust in them if they openly acknowledge using such a tool. Kurtis ^(talk) 03:30, 6 March 2026 (UTC)

A lot of editors who attempt something like what I've described would probably opt against doing so, as it would undermine trust in them if they openly acknowledge using such a tool.

I mean, it also undermines trust to learn about something the community strongly recommends doing (I realize LLMDISCLOSE is not really publicized that well) and decide "no, I'm not going to do that." It usually undermines it more -- generally if people are sanctioned for using AI it's because they are evasive about it, meanwhile there are a couple of people who are open about their process with AI who are not. Gnomingstuff (talk) 14:34, 6 March 2026 (UTC)

Well, yes—but that's assuming the ones who hide it get caught. If they don't, then nobody would be any the wiser. If someone believes that they won't get caught, then the only other incentive to be open about it becomes their moral compass. Kurtis ^(talk) 09:36, 7 March 2026 (UTC)

Yeah, shockingly, if you have no moral compass and make a deliberate effort to harm Wikipedia, sometimes you can get away with it. That's nothing new. Wikipedia's whole existence is built on the idea of people choosing to voluntarily cooperate with each other, including following community norms even when no one is watching. -- LWG ^{talk (VOPOV)} 15:27, 7 March 2026 (UTC)

WP:LLMDISCLOSE is an essay. Nobody has to follow it, it's optional. Cambalachero (talk) 16:05, 6 March 2026 (UTC)

Correct that no one has to follow it, but undisclosed LLM use will undermine trust with many people in the community, which is what the essay WP:LLMDISCLOSE is trying to help editors understand. On the other hand, disclosure is a good way to WP:DGF: Conversely, clumsily using an LLM in a transparent manner, promptly receiving relevant feedback, and responding reasonably to that feedback, would generally mean that the user is able to receive the message, following which they are just expected to improve their editing, motivated by what is in Wikipedia's best interest. -- LWG ^{talk (VOPOV)} 16:34, 6 March 2026 (UTC)

I have tried this before on a couple of articles and it turned out to be more work than just writing from scratch. So I would find it problematic for an editor to do this regularly because once you have built the writing muscle, that is so much easier than using the chatbot. 📎 JackFromWisconsin (talk | contribs) 19:09, 5 March 2026 (UTC)

Problematic? not necessarily but in practice this will both never happen or be done by people who have no idea what they're doing in the first place. Only someone with a decent amount of experience editing would be able to do this and those people don't need an LLM to write for them mghackerlady (talk) (contribs) 15:20, 9 March 2026 (UTC)

If you use AI to write an article, then fact check it thoroughly and rewrite sections to make it sound like it's not AI, it's literally your own work at that point. Obviously if it's accurate, cited information and it reads well, it's not a problem. The situation you describe is basically just using the AI as a structure/template and then writing an article using it. You're fine. Sentimental Dork (talk) 16:38, 9 March 2026 (UTC)

I should probably clarify that the above hypothetical was devised for demonstrative purposes. The fundamental question is: if someone admits to using generative AI for any facet of content creation, but doesn't specifically use it to actually generate content, is that still something that we should explicitly refuse to tolerate? Kurtis ^(talk) 18:07, 5 March 2026 (UTC)

Opinions vary on that point, but as EEng said, the non-hypothetical reality is that few to no examples of LLM use like you describe have been seen in the wild so far, while the LLM text we are actually seeing overwhelmingly fails verification. -- LWG ^{talk (VOPOV)} 19:38, 5 March 2026 (UTC)

There was an RfC about this topic only a few months ago. Gnomingstuff (talk) 19:40, 5 March 2026 (UTC)

Should something like the hypothetical scenario described ever actually come up… we can always look into it and decide on a case by case basis.

My one thought is that the editor in question made their job much harder by using the LLM… by simply summarizing the source material in their own words (from the get go) they could have saved themself a lot of extra steps (extra time and effort). Blueboar (talk) 20:37, 5 March 2026 (UTC)

• This reads like ragebait. Why would adding fact-checked, human-written material be problematic? (But the bait has inevitably been taken).Wh1pla5h99 (talk) 00:38, 6 March 2026 (UTC)

  To be clear, please: what reads as ragebait? EEng 02:37, 6 March 2026 (UTC)

  The whole post has the whiff of "let me act as if I need help tying my shoelaces". Wh1pla5h99 (talk) 02:45, 6 March 2026 (UTC)

  If an LLM is involved in any part of the process, it could cast the integrity of the editor into doubt in the eyes of many, even if they did their due diligence in ensuring that what they add is quality content. Kurtis ^(talk) 03:33, 6 March 2026 (UTC)

  I don't know why you would be worried about that. Wh1pla5h99 (talk) 04:45, 6 March 2026 (UTC)

  If an LLM is involved in any part of the process, it could cast the integrity of the editor into doubt in the eyes of many

  That's something they need to own, then. Gnomingstuff (talk) 14:29, 6 March 2026 (UTC)

• For things that involve actual writing, genAI's a net negative and I agree with EEng's scorched earth philosophy. But I think it's worth pointing out that a tool linked on WP:ATODAY was created using Claude AI, and having used that tool personally, it's pretty good. This is obviously not what you mean by genAI editing, but it uses genAI and it's for editing Wikipedia. It's worth stating the obvious that a tool like this can also be created without the assistance of AI, and I wouldn't be surprised if it would've been easier to do so. Tessaract2^Hi! 02:29, 6 March 2026 (UTC)
• I am curious. Two editors cite the high proportion of LLM generated text that "fails verification", and one example cites the WikiEdu process where we encourage newbie students to edit Wikipedia. What proportion of non-LLM-generated text on Wikipedia also "fails verification" (putting, presumably, our "I'm going to be strict" hat on). And what proportion of newbie student edits either fail verification or could be described as "hallucinated" or "clearly doesn't have the first clue" or "wouldn't make that mistake if they were an experienced editors like we all are".

To pick an article I hoped wouldn't be AI generated, I chose the User:Gnomingstuff's Agnes Grozier Herbertson. The first fact "(c. 1875 – 1958)" I'm going to guess appears somewhere in the 462 pages of the "The Oxford Companion to Edwardian Fiction". The citation lacks a page number. The second fact "a Norwegian writer and poet" I can characterise as "fails verification". A source cites Norway as her place of birth, but having Scottish parents, growing up in Scotland and later living in Oxford and Cornwall, and writing English-language works, I'm pretty sure there are no serious reliable sources describing her as "a Norwegian writer and poet". Our WP:MOSBIO says not to describe someone's nationality by their place of birth.

If I ask CoPilot "is Agnes Grozier Herbertson a Norwegian writer and poet" it replies: "Yes — Agnes Grozier Herbertson was a Norwegian writer and poet. According to multiple reliable sources, she was born in Oslo (then Christiania) and is consistently described as a Norwegian writer and poet, even though she later lived in the United Kingdom." The so-called "multiple reliable sources" include the Wikipedia article and another that appears to have lifted some text from the Wikipedia article. This makes me reasonably sure that falsehoods that fail verification predate the AI plague on Wikipedia, and that our fallible human editors are themselves the source of a great deal of AI's incorrectness. -- Colin°^Talk 14:36, 6 March 2026 (UTC)

Please don't ping me for discussions I am already aware of. The article isn't AI-generated, thanks for pointing out the style and sourcing discrepancies.

That being said, "fails verification" in the context of AI generally means that AI attaches a source to text when the source does not back up that text, and sometimes may not say anything even remotely like it. Generally this suggests that whoever added it didn't even bother reading the source (considering that many such issues can be found in 5-10 minutes' worth of spot checking). It doesn't generally mean unsourced content. Gnomingstuff (talk) 15:17, 6 March 2026 (UTC)

EDIT CONFLICT - basically what the unpingable user above said: There is no source attached to the claim of Norwegian nationality there, so this is not a source-text-integrity issue. No one is claiming that humans are infallible, what we are claiming is that AI routinely places in-line citations on text that those citations do not verify, and that a human who read and comprehended the source would be unlikely to think they did verify. -- LWG ^{talk (VOPOV)} 15:29, 6 March 2026 (UTC)

Text on Wikipedia "fails verification" regardless of the "context" of who or what placed it. Students on the WikiEdu programme (and newbies generally) have been placing text that fails verification, and attaching sources to text that do not remotely back up the text (verify it), long before AI came along to help them. LWG claims there is not a source-text-integrity issue in the example. Lead sentences routinely lack citations and we assume and hope verification is fulfilled by the body. The body contains several "life and works" citations that would be expected to verify the claim that Norway holds this writer and poet as one of their own. This false lead claim and all the body citations were added in the first edit. So, that's a human doing exactly what you claim is an identifying failure of AI. Someone/something created some text that cites some sources, and it turns out at least one claim has no basis within them, nor indeed, within any reliable sources. The only difference here from "some bollocks a random person on the internet wrote" to "an AI hallucination" is because you guys are keen to frame an age old issue in a novel way.

When we are prejudiced, we see only problems and limitations. There is a "human hands have five fingers" problem with some of the ability of AI right now, coupled with a "newbies aren't very good at prompting and using AI only for what it is currently good at". The Wright brothers flew 260 metres and for less than a minute. Sixteen years later, someone crossed the Atlantic non-stop. Sixty-six years after that first flight, we landed men on the moon and brought them back safely. And we did that with slide rules and spanners.

We have here a question asking, really, what would satisfy as an acceptable usage of LLMs in editing Wikipedia. And for some here, there isn't an acceptable use ever, and to justify that, all the current limitations are presented as though specific to AI (most are not) and without consideration, really, that the Wikipedia created by humans has lots of flaws one could enumerate too. If one picks random text, one can find all sorts of problems, so it isn't at all surprising to me, that if one picks AI generated text, one can find all sorts of problems.

I have no doubt that in a short period of time, we will routinely pass our text through polishing by AI. We will routinely get AI to locate and summarise sources and draft material. We will so routinely use AI to do text-source verification that the GA and FA processes will make it a requirement that nominators have done so before wasting human time. Editors will use AI to seek out dodgy sentences and articles for deletion. Edits by newbies will be vetted by AI before appearing in the text. And so on. You guys are looking at the Write brother's plane and scoffing that it just lifted off 3 metres from the ground for 59 seconds. Maybe it would be more productive if we put our efforts towards using it better, rather than fighting it. These debates will look frankly embarrassing in a short space of time, and be cited humorously in mocking listicles. -- Colin°^Talk 20:29, 8 March 2026 (UTC)

That's a lot of words to say "don't be a fuddy-duddy, the technology will improve until it becomes useful, trust me bro". I'm more concerned with the present reality than the hypothetical future. I'm still open to being linked to any examples you have of AI being used to improve the wiki so I can explore new use cases, and I'd encourage you to check out the examples I linked below to see the kinds of issues that arise in LLM text, and then pick up a shovel and head over to WP:AINB to help us work through the backlog of bad LLM content that has already been identified. -- LWG ^{talk (VOPOV)} 01:14, 9 March 2026 (UTC)

There are many editors who see lots of bad content (they suspect to be) generated by LLMs and extrapolate that to "LLMs will always be bad, trust me bro". That attitude is exactly as unhelpful to everything as the "LLMs will inevitably be great in the future" comments.

The reality is between the extremes - it is currently easier to produce bad content with LLMs than it is to produce good content with LLMs, but producing good content is not impossible (see the examples given in the numerous other discussions, and also note that most good content produced by/with the aid of LLMs will go unnoticed because almost nobody checks good content to see if LLMs were used or not in the same way that neutral content about notable subjects written by undisclosed paid editors is not flagged). Some editors regard the time spent polishing LLM output as worthwhile. Unless you are paying that person for their time it's none of your business whether you agree with them or not.

It is probable that LLMs will get better. How much better and on what timescale is unknowable, but it is possible (although I personally think it unlikely) that it will become the panacea some people hope for. Our policies and guidelines should focus on the outcomes we want - i.e. well-written content verified by reliable sources - without regard for the the method used to create it, because as long as it's not a copyright violation the method is completely irrelevant. Thryduulf (talk) 01:31, 9 March 2026 (UTC)

Still waiting on that apology for saying I should be blocked. If you're so concerned about policies and guidelines then follow them. Gnomingstuff (talk) 04:03, 9 March 2026 (UTC)

I have never said you should be blocked. Thryduulf (talk) 10:35, 9 March 2026 (UTC)

Your response to my simply... participating in AI cleanup work... was it's amazing you've not been the subject of a CIR block.. Which you have not apologized for. Gnomingstuff (talk) 14:06, 9 March 2026 (UTC)

Please refrain from selective quotation. What I actually said was if you cannot understand the difference between a content dispute and the general issues with content that get discussed here then it's amazing you've not been the subject of a CIR block. As you haven't been, then I can only presume that your comments here are not the result of incompetence but something else.

I hope and assume the explanation for your misleading quotation and false accusation is neither bad faith nor lack of competence. Thryduulf (talk) 18:40, 9 March 2026 (UTC)

Sorry, no. Saying "it's amazing you haven't been the subject of a CIR block" is, implicitly, claiming that someone should be blocked. Otherwise you'd say something like "it's amazing that someone has not mistakenly applied a CIR block," or something else that implies you do not personally endorse that. I don't know what the fuck you have against me. Gnomingstuff (talk) 19:08, 9 March 2026 (UTC)

If you actually read the whole of what I actually wrote, rather than reacting to something I didn't write then it should make complete sense. Seriously, read the whole thing, especially including the final sentence.

I don't know whether it will help but, I can try expressing it terms of semi-formal logic:

• A happened.
• B is one plausible explanation for A.
• If the explanation for A is B, then C would be the consequence.
• However C has not happened.
• Therefore B cannot be the explanation for A.

Nowhere in that is there any suggestion that C should happen if the explanation is not B. Thryduulf (talk) 21:05, 9 March 2026 (UTC)

So if I understand correctly, your only purpose in that comment was to say "Gnomingstuff is clearly competent, as evidenced by that fact that he hasn't been CIR blocked?" What was the point of even saying that? -- LWG ^{talk (VOPOV)} 22:09, 9 March 2026 (UTC)

To understand the point of the comment you will need to read the whole exchange in its original context. I'm not repeating it here because it's really not relevant. Thryduulf (talk) 23:41, 9 March 2026 (UTC)

Thank you for saying that. I agree that what matters is the quality of the content and sourcing in articles, and that worrying about how the content and sourcing were generated is a distraction. I am firmly opposed to any attempts to prevent me from using LLMs as aides in creating and maintaining content. The content and sources I add to Wikipedia should be judged on their adherence to policies and guidelines, and not at all on whether I used an LLM. Donald Albury 12:35, 9 March 2026 (UTC)

Thryduulf, I respect your years of service to this project, but your behavior on this topic has become WP:BADGERING. If you would like to engage with my actual position, I'd be glad to talk with you. If you are going to use my comments as a jumping off point to repeat your same position and throw shade at editors who disagree with you, I would respectfully ask you to stop replying to me. In case it's still unclear after all the previous discussions, my position is: LLM-assisted editing should be permitted, but disclosure should be required and WP:CIR should be enforced, with the WP:ONUS falling on the editor to demonstrate that they know what they are doing and that their LLM-assisted contributions are constructive. -- LWG ^{talk (VOPOV)} 14:27, 9 March 2026 (UTC)

If you wish to accuse me of bludgeoning every time I make a comment feel free but don't expect me to agree with you. If you wish to assume that my comment was a personal attack then that's your prerogative (I can assure you it wasn't) but it's a bit rich to simultaneously berate people for not engaging with then substance of your comment when (a) I did, and (b) you didn't extend that courtesy to me. Thryduulf (talk) 23:46, 9 March 2026 (UTC)

STOP - and this is directed at more than one editor! While I don’t think anyone intended their comments to be personal attacks, this conversation is getting personal enough that others are taking it that way. I suggest that everyone take a break from this thread. You have all made good arguments to support your respective views on this issue. Let it rest there. We don’t “win” by being the last to respond. Blueboar (talk) 00:14, 10 March 2026 (UTC)

Yes please. I accidentally subscribed to this topic, so I keep getting excited about new notifications and every single time it turns out that it's just these guys arguing more. Sentimental Dork (talk) 00:28, 10 March 2026 (UTC)

–For comparison, here are articles generated from those sources using Grok and using ChatGPT, so you can compare and see how the source use and style issues differ. -- LWG ^{talk (VOPOV)} 15:50, 6 March 2026 (UTC)

Maybe I'm misunderstanding, but I'm seeking two pages listing prompts you used but not the results that you got from them. -- Nat Gertler (talk) 17:52, 6 March 2026 (UTC)

Oops, I accidentally linked to the talk tab rather than the content tab. The LLM output is at User:LWG/AISandbox and User:LWG/AISandbox2, and the prompts are described on the corresponding User_talk pages. -- LWG ^{talk (VOPOV)} 17:56, 6 March 2026 (UTC)

Makes sense. -- Nat Gertler (talk) 01:05, 7 March 2026 (UTC)

Of course not. The only thing not allowed is posting AI-generated content. Use it to help you as much as you want. Sentimental Dork (talk) 19:25, 9 March 2026 (UTC)

Recommended reading: https://wikiedu.org/blog/2026/01/29/generative-ai-and-wikipedia-editing-what-we-learned-in-2025/

Their experience was that chatbots were bad at creating articles, but that they could be useful. For example, instead of producing sentences/paragraphs, you could probably get an outline that you turn into prose yourself, or perhaps with some LLM tools, a list of key facts, paired with sources that you could then check (e.g., "Birth: 32 Octember in Novia Scotia. Source https://example.com, which says "Alice was born in Novia Scotia on Octember 32nd" – and then you could look in the source to see whether the quoted text is actually in it). WhatamIdoing (talk) 23:39, 10 March 2026 (UTC)