Wikipedia:Administrators' noticeboard/IncidentArchive1125

User 2600:1702:3650:9050:88E3:9399:5608:8A92 (talk · contribs), with no edits before today, has replaced the entire lead in eight biographical articles of French statesmen or politicians in 28 minutes. Maybe we need to welcome them with a special barnstar created for the occasion, and instant adminship! Or, could it be...<spooky font>Chat GPT?</spooky font>

IP replaced the lead of Charles de Gaulle on 7 April at 01:52 (diff). In my opinion, (read: "I have not the slightest effing tiny little doubt in my mind") these have all been generated by Chat GPT. (I've been experimenting a fair bit, and I recognize the style pretty easily, now.) The CDG lead is pretty good, but it doesn't necessarily summarize the most important points in the body of the article, and I'm not convinced everything in it is already covered in more detail in the lead, where references can be found.

I've raised a discussion at Talk:Charles de Gaulle/Archive 2#Replacement of the entire lead on 6 April by IP user via LLM, reverted IP's edit at the CDG article, and reverted several other of their edits (as have Manticore and Galobtter; thanks!):

(So far, nothing else of interest in 2600:1702:3650:9050:0:0:0:0/64.)

Finally, the larger issue of LLM use at WP is an unsettled one at Wikipedia, and I'm aware of the growing body of discussion both here and elsewhere about it (see the tip of the iceberg here), so I thought this should be brought to the community's attention as soon as possible.

For starters, a question: I plan to revert any edit by this user that smells of LLM to me, and I'm getting pretty good at spotting it, but that's a pretty iffy reason, policy-wise. I trust myself, but you don't necessarily need to agree, and if there's a good reason not to follow my plan, please weigh in. The more important question, of course, is the more general one. This may also be of interest to @Fuzheado, CactiStaccingCrane, and Mztourist:; feel free to add others. Thanks, Mathglot (talk) 06:26, 7 April 2023 (UTC)

I entirely agree with User:Mathglot's position. Mztourist (talk) 06:32, 7 April 2023 (UTC)

I read two of them, Charles de Gaulle and Nicolas Sarkozy. These edits are summarizing the summary and significant introductory content is lost from the leads. This is automated dumbing down and a clear degradation of our content. I favor an immediate indefinite block from article space for any editor who pulls a stunt like this. Cullen328 (talk) 06:45, 7 April 2023 (UTC)

Hear hear! I looked at what happened to Mitterrand: replacement of nearly all the factual summary by bland grammatically-correct blurb that could have been written about anyone from the pope to the florist at the end of my street, the sort of deceptively-correct cosmetic-English that AI is good at producing. There is no way it was an improvement; there is no way AI can improve in a situation like this, as it must either summarise (which had already been done) or look beyond the text (in which case it needs to support itself with referencing like any other editor). But fundamentally, if a person is capable of writing half-way decent English they don't need AI to rewrite leads, and if they aren't capable of writing half-way decent English, they can't assess the AI's writing. And we need honesty from our editors: if they cannot admit they aren't a human, what editorial ethics do they have? Elemimele (talk) 07:05, 7 April 2023 (UTC)

A-Admit I'm not a human?... I'm definitely a human, beep boop. LilianaUwU ^{(talk / contributions)} 07:12, 7 April 2023 (UTC)

Yet another reason to ban all LLM use on Wikipedia. LilianaUwU ^{(talk / contributions)} 06:59, 7 April 2023 (UTC)

I think this definitely furthers the case for banning LLM use on Wikipedia. Even though the edits were fairly obviously detrimental, and completely changed the leads of high profile leads, they still took hours to revert - I can only imagine how much easier it would be sneak in referenced content that looks perfectly reasonable but is totally made up/full of AI hallucinations and factual inaccuracies. Galobtter's sock (talk) 07:09, 7 April 2023 (UTC)

Banning LLMs might be useful in dissuading already established editors from using them, but what would that do when it comes to IPs like this one? They're not exactly well-versed with our policies, and most don't care to be. PopoDameron ⁠talk 10:57, 8 April 2023 (UTC)

FWIW: based on their editing history, this editor is clearly User:President of Switzerland. They were blocked for 31 hours last month, see User talk:2600:1702:3650:9050:E969:B1EA:947:E5C5. This recent attempt at “summarization” involved replacing the intro with verbatim quotes from the body. 100.36.106.199 (talk) 10:45, 7 April 2023 (UTC)

I'm less worried about the lead rewrites, which are not too hard to spot, than with the ability to fake sources, or to generate fiction with real sources. Consider this snippet for example, and tell me what you see:

Clotharperic was a Frankish king who reigned over a kingdom that covered much of modern-day France and Germany in the late 6th century.<ref>{{cite book |author=Wood, Ian |title=The Merovingian Kingdoms: 450-751 |publisher=Pearson |location=Harlow |date=1994 }}</ref> He was the son of King Clotaire I, and upon his father's death, he inherited a significant portion of the Frankish kingdom.<ref>{{cite book |author=Wallace-Hadrill, J. M. |title=The Long-Haired Kings, and Other Studies in Frankish History |publisher=Methuen |location=London |date=1962 }}</ref> Clotharperic was known for his military conquests, having expanded his kingdom's borders through successful campaigns against neighboring peoples.<ref>{{cite book |author=Wood, Ian |title=The Merovingian Kingdoms: 450-751 |publisher=Pearson |location=Harlow |date=1994 }}</ref> He was also a patron of the arts, and under his reign, many magnificent works of art and literature were created.<ref>{{cite book |author=James, Edward |title=The Franks |publisher=Blackwell |location=Oxford |date=1988 }}</ref> Despite his successes, Clotharperic was not without his detractors, and his reign was marked by political turmoil and assassination attempts.<ref>{{cite book |author=Wood, Ian |title=The Merovingian Kingdoms: 450-751 |publisher=Pearson |location=Harlow |date=1994 }}</ref> Nevertheless, he remains a significant figure in Frankish history, and his legacy continues to be studied and debated by historians today.<ref>{{cite book |author=Wallace-Hadrill, J. M. |title=The Long-Haired Kings, and Other Studies in Frankish History |publisher=Methuen |location=London |date=1962 }}</ref>

Chat GPT turned the tables on me, because I was expecting a mix of real and faked references, as I've mostly seen in the past, but these are all real (usually I ask for a lot more than one paragraph, so maybe that's why). The interesting thing here, is that king Clotharperic is a figment of my imagination (zero hits in Google, but that may change by the time you read this ), and Chat GPT "sourced" it anyway with real references. If I had picked a real king, and the bot gave real references, who is going to check them? This is what really worries me, because if an editor doesn't identify bot usage in an edit summary, I just don't know how we're going to discover it. Mathglot (talk) 09:26, 7 April 2023 (UTC)

You may be interested in WP:ANI#Suspected hoax content and LLM use by User:Gyan.Know. Which was about ChatGPT and fake referencing. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 10:50, 7 April 2023 (UTC)

Here is what concerns me: We can unequivocally ban the use of ChatGPT or LLM on Wikipedia. That is the easy part. Writing a policy, having the community agree upon it, is comparatively easy to what will be the real challenge - enforcement and identification. The IPv6 showed its hand pretty early by hitting all eight of the aforementioned articles within half an hour, but we will have a harder time sussing out those edits that are introduced with more stealthy or considered approaches that I won't expound upon due to WP:BEANS. And any editor that is determined enough won't give a damn what our written policy is¹. Moreover, as the software improves (and it will improve), the line is going to increasingly blur to where we run the risk of friendly fire against good-faith editors whose edits may bear a passing resemblance to ChatGPT. So far our ability to discover AI edits is based on intuition and pattern recognition, all forms of inductive reasoning and, as Mathglot says, very little deduction due to lack of prima facie evidence. Because of that, I wonder what a "no LLM editing" policy accomplishes other than just expressing our collective outrage. --⛵ WaltClipper -(talk) 14:32, 7 April 2023 (UTC)

Footnote 1: One alternate solution is to outright ban IP editing on Wikipedia as a result of the heightened risk of LLM-editing created by having such a wide attack vector, but I don't see this happening ever, and certainly not for this reason. --⛵ WaltClipper -(talk) 14:35, 7 April 2023 (UTC)

I've been advocating that for years, but no matter how much evidence and argumentation is put forward that doing so would be beneficial to the project, it doesn't seem as if it's ever going to fly here, unfortunately. Beyond My Ken (talk) 19:21, 7 April 2023 (UTC)

Detection is definitely a big problem. But I think a "no LLM editing" policy would at least make sure that good faith users know not to use LLMs to generate text for articles - and we have always relied on there being a lot more good-faith editors than malicious ones. Right now I'm seeing a wishy-washy approach from the community (at least the one that is outlined at WP:LLM) that doesn't seem to acknowledge the risks involved, and seems to think that humans are actually going to meaningfully verify what the LLM writes (which would take longer than writing the text by yourself).

(as a side note, I just saw this article about how ChatGPT invented a sexual harassment case against a professor - scary stuff.) Galobtter's sock (talk) 18:51, 7 April 2023 (UTC)

I think an expression of collective outrage could be a good thing. If the unethical products of a toxic industry are going to drag down our attempt to use the Internet to educate humankind, we might as well be angry about it. XOR'easter (talk) 19:05, 7 April 2023 (UTC)

And yes, WP:LLM was written in a fantasy land where the only people adding LLM content to Wikipedia will be those who are so conscientious about double-checking and revision that they could start their rough drafts by throwing the bloody I Ching and still end with a respectable article. XOR'easter (talk) 19:14, 7 April 2023 (UTC)

Yes, in fact I'd say almost by definition those using LLMs to create content are those least likely to (or to even be capable of) double check the content. Galobtter (pingó mió) 08:06, 8 April 2023 (UTC)

I once uncovered a decade-old WP:HOAX biography. It was well-written and entirely plausible; but contained one tiny error which may not even have been such when it was posted: it linked to a DAB page. The problem was that I could find no evidence that several of the key references (all print-only, no later than C19, and in an accessible language) existed. My emotions after completing the investigation were mixed: satisfaction at a job well done, admiration for the hoaxer's skill, and something approaching fury that I wouldn't get that hour-and-a-bit of my life back. The thought of mass-produced articles of that kind fills me with horror. Narky Blert (talk) 01:20, 8 April 2023 (UTC)

The ultimate answer to that may wind up being significantly tightening the minimum requirements for article creation. Not at all what most veterans would prefer, and not within the scope of ANI -- that'd be a site-wide RfC -- but this is downright an existential threat to the encyclopedia. All we need is for the media to start trumpeting how Wikipedia's being written by chatbots, and the encyclopedia's credibility is gone for good. Ravenswing 01:41, 8 April 2023 (UTC)

I don't have a copy, but from memory More Random Walks in Science ed. R. L. Weber (1982) contains two fictional biographies which made their way into a printed encyclopedia.or compendium, whose editors adopted a very stuffy attitude when they found out they'd been taken in (compare the Sokal affair). They were of Jean-Marie Litre (a simple-minded French aristocrat executed during the Reign of Terror, whose name is commemorated in the litre) and of Luis Herrera Torsión (a Spaniard tortured by the Inquisition, whose sufferings gave their name to the torsion balance). They were very obviously practical jokes to anyone who actually read what they said, rather than being written to deceive. Narky Blert (talk) 06:25, 8 April 2023 (UTC)

@Ravenswing: I agree about a site-wide Rfc (possibly with some preliminary discussion, to narrow the Rfc question before we start), and I don't think we should wait too much longer to attempt it. How do you feel about WP:VPP as a venue? Mathglot (talk) 11:35, 8 April 2023 (UTC)

That's certainly where it should start, I believe. Ravenswing 13:55, 8 April 2023 (UTC)

I'd be happy to contribute; feel free to {{ping}} me. Narky Blert (talk) 17:18, 8 April 2023 (UTC)

Agreed, please also ping me. But – but – I know we'll be backing a losing cause when it comes to consensus. At my last unscientific check, ⅓ of editors who expressed an opinion don't think that AI/LLM is a problem at all, ⅓ of editors think AI/LLM is useful, and ⅓ of editors wish the issue would go away until they've had time to think about it properly. Add that to the current 60:40 inclusionist:deletionist ratio and we're on a hiding to nowhere. Nevertheless, even some doomed battles are worth fighting. — Trey Maturin™ 18:38, 8 April 2023 (UTC)

I think that's much too pessimistic, especially as pro-ban editors are more likely to comment than anti-banners, and the issue has received a lot of publicity recently. We also need a process for reverting suspected AI additions. Do the sites produce the same content when asked twice? Johnbod (talk) 18:54, 8 April 2023 (UTC)

It's received a lot of publicity recently, but we've still, on this page right now alone, had two otherwise unobjectionable established editors go to ChatGPT, type in a question, post the answer to a new article here and then not understand at all in any way why this was a problem – despite the disclaimers and the publicity and the news articles and everything else. They thought a chatbot was 100% spot-on Wikipedia material. And then a third of the people responding to this had no problem with this but were unhappy with the editors being blocked from mainspace because they meant well. We're on a hiding to nothing in trying to get consensus. Even the no-AI-please warning templates ({{uw-ai1}} etc) got pushback from established editors saying that this "deletion on sight" policy was bad and they would prove so by posting AI/LLM articles to mainspace themselves. — Trey Maturin™ 19:34, 8 April 2023 (UTC)

@Johnbod They shouldn't, no. They will tend towards producing similar content, but unless they start caching ChatGPT and BingChat answers for popular questions (I don't believe they do this, but perhaps I am wrong, in which case I would appreciate being corrected!), the answers are newly generated each time. Using specific phrases in "prompt hacking" can duplicate weird hacks to a certain extent, but not typically exactly. -- asilvering (talk) 00:51, 9 April 2023 (UTC)

Which makes it a thousand times harder because it means there's no common phrases or giveaways we can use to filter out the AI-written content (yet, anyway; if we're lucky there's certain word strings that the AI likes to use that might make it possible to flag its writing, but I suspect we would need a lot more data to see if that's how it works.) StartGrammarTime (talk) 01:34, 10 April 2023 (UTC)

Even if we find some strings that current models are more likely to output, within a couple of years there'll be completely different models that don't behave the same way, so it would be a very temporary solution. PopoDameron ⁠talk 02:41, 10 April 2023 (UTC)

At this rate, it's more like "within a couple of months". -- asilvering (talk) 03:40, 10 April 2023 (UTC)

• Editor MVM (talk · contribs · deleted contribs · logs · filter log · block user · block log). This user is constantly adding promotional/glorifying content to articles related to Nair Service Society, despite multiple warnings.

jojo@nthony (talk) 15:28, 9 April 2023 (UTC)

Editor MVM also moved an article in order to add honorifics to the title. MVM hasn't responded to any of the messages on their talk page. Perhaps an admin could block them from article space until this editor engages in discussion? Schazjmd (talk) 15:42, 9 April 2023 (UTC)

 Done Blocked from articlespace until they start engaging. Black Kite (talk) 18:23, 9 April 2023 (UTC)

It's a bit late to point this out but for the record, the editor in question has still not been notified about this discussion. I've notified them. Something something big red box something something Burma Shave. GabberFlasted (talk) 11:34, 10 April 2023 (UTC)

Someone, or multiple people seem to have a desire to create a Wikipedia article about the Bihar Engineering University in Patna.

There's currently 2 duplicates at BEU Patna and Bihar Engineering University, Patna, and Bihar Engineering University has been repeatedly deleted. There's also two drafts at Draft:Bihar Engineering University and Draft:Bihar Engineering University 2.

It appears that User:EG2009TU, User:Amarjeetkr97, User:R7461867442 are collaborating, or WP:SOCKPUPPETS, to create an article on this university. Also, User:GargAvinash has edited one of the drafts.

I'm not an expert of sockpuppetry or editors who are determined to create an article, despite all opposition, so I'm asking for advice from admin(s).  – CityUrbanism 🗩 🖉 14:06, 9 April 2023 (UTC)

Bihar Engineering University has been salted against recreation, after being moved to draftspace by User:Liz. I've just salted Bihar Engineering University, Patna myself, and prodded BEU Patna as non-notable. The only remaining potential problem now should be that there are two drafts, but I'm not sure that matters, as long as they don't emigrate to mainspace again. And of course there remains the possibility of creating yet other spellings (I won't suggest any examples, per WP:BEANS). And yet other socks.. sigh. As for obsession, such obsessions usually come about because of paid editing. It all comes with the territory, I'm afraid. Thanks for watching, CityUrbanism. If you should happen on more versions, feel free to contact me directly. Bishonen | tålk 13:22, 10 April 2023 (UTC).

Please see: "https://en.wikipedia.org/wiki/User_talk:47.232.88.19" at the bottom. Can someone please tell me if this guy is right and not just out to get me and if he is right specifically why he is right? He just posts links to entire articles as an explanation but not what part of it he thinks I am not following. Totally unfair and uncool. As far as I know I am totally within the rules here and I was totally on topic. I was literally agreeing with something someone else wrote on there and my comment was basically similar to another comment posted there but that guy got to keep his comment and I was not. This is totally unfair! And then he gets cross with me when I ask for explanation. — Preceding unsigned comment added by 47.232.88.19 (talk) 16:31, 29 March 2023 (UTC)

Also, sorry if wrong place to complain, please let me know where to file the complaint if this is wrong place. 47.232.88.19 (talk) 16:37, 29 March 2023 (UTC)

People just love taking me to ANI You've failed to notify me of this discussion per the bright red notice at the top. While that's technically moot now since I'm aware of this discussion it's still something you should do. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 16:40, 29 March 2023 (UTC)

I did. Read the comments at the bottom of the thing. I told you that I posted it here and even provided a link 47.232.88.19 (talk) 17:59, 29 March 2023 (UTC)

Maybe you should actually read it a bit more carefully. The notice says "When you start a discussion about an editor, you must leave a notice on the editor's talk page.". You did no such thing. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 18:01, 29 March 2023 (UTC)

I did! Look at my talk page where you were scolding me. I directly replied to you with a link! 47.232.88.19 (talk) 18:05, 29 March 2023 (UTC)

How about I actually bold the relevant part for you. "When you start a discussion about an editor, you must leave a notice on the editor's talk page." Not the person who is reporting's talk page, the user who is being reported's talk page. You never gave me any notice on my talk page. In fact, I didn't have any talk page notice regarding ANI today until User:Tails Wx came along and added it (only to revert themself since I was already aware). I think you need to take some time and read things completely so you actually understand what you are being told. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 18:10, 29 March 2023 (UTC)

Oh come on. Cut me a break. Clearly I was trying to notify you, but apparently misunderstood the word editor to mean myself. 47.232.88.19 (talk) 18:17, 29 March 2023 (UTC)

Did you not read what Blaze stated? When you start a discussion about an editor, you must leave a notice on the editor's talk page. Pings aren't sufficient, see this discussion. Tails Wx 18:32, 29 March 2023 (UTC)

I think the IP understands that, judging by what they're saying --Licks-rocks (talk) 18:48, 29 March 2023 (UTC)

Mk. Just a clarification, to the IP, I'm sorry if I was a bit bite-y. Tails Wx 19:02, 29 March 2023 (UTC)

The difference between ScottishFinnishRadish's comment and your comment was that ScottishFinnishRadish was discussing the article, and you were discussing Mr. West. ScottishFinnishRadish's comment was made in December, and was discussing the categories at the bottom of the page. If you check the current categories, Mr. West is no longer included in Category:American neo-Nazis. You came along 3 months later and made a comment that did not address what should or shouldn't or is or isn't in the article, rather expressing your thoughts on West as a person. That's the difference. ~ ONUnicorn^{(Talk|Contribs)}problem solving 16:54, 29 March 2023 (UTC)

I disagree with this characteriazation. I was basically repeating what he said. IT wasn't even my own opinion. I was seconding someone else's. 47.232.88.19 (talk) 18:25, 29 March 2023 (UTC)

And the page actually still says he identifies as a nazi. 47.232.88.19 (talk) 18:26, 29 March 2023 (UTC)

Wait. If a topic is still kept up does that mean it could still have possibly been solved? I thought it meant it was still an open discussion. 47.232.88.19 (talk) 18:30, 29 March 2023 (UTC)

While a discussion that hasn't been closed properly (via the various methods at WP:CLOSE) can still be replied to, if there haven't been any responses in a while (usually around 7 days/1 week for most things) then it can be safe to say that the editors have moved on and there's no need to reply. If you think there's a good reason for you to make a comment feel free to do so (there's technically no official policy or guideline against necroposting, tho I myself don't see it as very constructive depending on how long ago the last reply was) so long as it isn't actually closed. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 18:37, 29 March 2023 (UTC)

Yes, Blaze Wolf was correct. Please read WP:NOTFORUM Catfish Jim and the soapdish 16:41, 29 March 2023 (UTC)

Ah, but they've already "skimmed" that page, and their takeaway from it was that random forum-like comments are fine and tu quoque is a valid position to argue from. So that's us told. — Trey Maturin™ 16:50, 29 March 2023 (UTC)

• The IP is apparently unaware that the official WP position is "new editors can fuck right off". I'm sorry you went through this, IP. There are a lot of smug editors here who enjoy throwing their weight around, because they're too cowardly to quote "NOTAFORUM" to User:ScottishFinnishRadish. --Floquenbeam (talk) 16:56, 29 March 2023 (UTC)

  Thank you! And you were the person I was agreeing with by the way. But apparently me posting that I second your opinion is against some as of yet unknown rule. 47.232.88.19 (talk) 18:04, 29 March 2023 (UTC)

  I believe this was intended as sarcasm. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 18:06, 29 March 2023 (UTC)

  I am NOT being sarcastic. Sorry if it came across like that, but I was not being sarcastic. 47.232.88.19 (talk) 18:12, 29 March 2023 (UTC)

  I'm not referring to your comment. I'm referring to User:Floquenbeam's. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 18:15, 29 March 2023 (UTC)

  Hey but while you are here, can you actually explain, withOUT getting cross with me, why others who had similar comments were given a pass, but for me my first comment was immediately flagged and removed? You referenced a bunch of article on policy, but did not actually mention which part of those articles you thought I was in breach of. From what I read, I think I am within the rules. Also, I've recently encountered something called wikilawyering. I think that this whole thing counts as lawyering. A small comment being attacked with this much force? I mean I feel like you are basically telling me that I just can't participate period if I have to second guess every single thing I have to say. I wasn't disrespectful at all and was totally on topic! 47.232.88.19 (talk) 18:23, 29 March 2023 (UTC)

  @Blaze Wolf: It was sarcasm in the sense that we probably shouldn't act like new editors can fuck right off, but it was serious about you doing an extremely poor job of interacting with this editor, and Catfish Jim and Trey Maturin didn't help either. If someone had told the IP editor what ONUnicorn took the time to tell them, above, I doubt we'd be where we are now. Just because the comment wasn't helpful doesn't mean we need to edit war with them and then refuse to explain. Floquenbeam (talk) 18:26, 29 March 2023 (UTC)

  The IP was referred to the policy page in question, agreed that they had read it (and then said they'd skimmed it) and what they took away from reading/skimming the page was that they were right and should head to a drama board and attempt to get sanctions against the person who pointed them to the page. I'm sure we can simplify most of our policy pages down to a single paragraph in order to make them graspable, but should we? — Trey Maturin™ 18:34, 29 March 2023 (UTC)

  Pretend you don't know much about WP. Now follow this link: WP:NOTFORUM. You didn't do it, did you, because you think you know what it says, but humor me. It links to the middle of WP:NOT; a giant, sprawling catch-all page. The fact that it links to a specific section of that page is not obvious to a new editor; you have to know what the shortcut box means, and you have to know to look at the right margin to see it, you have to know how it applies to your situation. The section in question isn't highlighted, it's just floating there mid-page. Add to that the fact that, frankly, this is an edge case of NOTFORUM at best; you could make a case that it applies, and a case that it doesn't apply. It's OK to link to NOTFORUM at first, because there are only so many hours in the day. But when the IP politely said they didn't understand, they got nothing but attitude. We see hundreds - probably thousands - of comments like the IP's comment every day, many by more experienced accounts, and many actively harmful, instead of this harmless "me too" comment. Personally, I think it harms WP more than it helps to remove a harmless comment like this. I don't know why Malerooster chose to revert this particular comment, but at least they were polite about it and left a welcome template. I really don't know why Blaze Wolf chose to edit war with the IP about this one, and answered with more useless bluelinks and templated warnings. Established editors forget to notify people every day, but we're spending multiple paragraphs yelling at the IP they didn't do it. The IP did not ask for a sanction, but you're righteously indignant, convinced that they did. This is all just an example of the smug "new editors, especially IP editors, can fuck right off" attitude you're displaying. Floquenbeam (talk) 19:11, 29 March 2023 (UTC)

  That last sentence is unworthy of you, Floq. — Trey Maturin™ 19:21, 29 March 2023 (UTC)

  OK, struck, but that was like 5% of my post. And rude or not, unwise to say out loud or not, it is honestly how I feel about the response here. --Floquenbeam (talk) 19:53, 29 March 2023 (UTC)

  Perhaps I was a bit short and shouldn't have answered the question as put, as a binary choice (I agree, it did come across a bit WP:BITEY). I doubt Blaze Wolf is just out to get the IP. I would be happy to offer some mentorship to the IP. Catfish Jim and the soapdish 18:49, 29 March 2023 (UTC)

  Oh wait. I thought Scottish was the guy who wrote this comment. Scottish was the guy I was agreeing with. But still thanks regardless. 47.232.88.19 (talk) 18:11, 29 March 2023 (UTC)

• It seems like we have a WP:BITE problem here. IP47 was mistaken about how Wikipedia talkpages work, but instead of having it explained to them they were reverted and templated. I don't really blame BlazeWolf in this circumstance, I'm sure that page gets a lot of nonsense. But we should try to remember that West's page is probably a magnet for new editors especially now, and we should probably avoid reverting comments just because they're not fully relevant. Reversion of talk page comments should really only be for blatant trolling, vandalism, BLP vio etc. The Wordsmith^{Talk to me} 18:46, 29 March 2023 (UTC)

  Reversion of talk page comments should really only be for blatant trolling, vandalism, BLP vio etc. I'm not following. So you're saying we should keep talk comments that violate WP:NOT, which is policy, if it isn't obviously vandalism? ― Blaze Wolf^Talk_{Blaze Wolf#6545} 18:49, 29 March 2023 (UTC)

  If you really hate pointless comments that much you could've hatted it, we generally frown on deleting other people's talk page comments after all. But since it didn't get in the way of any ongoing discussion and the editor cannot be expected to be aware of any of the applicable parts of WP:NOT, and even apologised in their comment for not being aware of the rules ahead of time, the better approach here would probably have been to just leave it be and place a welcome template at the IP's talk page. --Licks-rocks (talk) 19:18, 29 March 2023 (UTC)

  I know that WP:NOT says "Material unsuitable for talk pages may be subject to removal per the talk page guidelines" (emphasis mine) but that doesn't mean that reverting and warning is always (or even usually) the best option. When a comment in a discussion seems to be made in good faith, or a new editor doesn't understand how talkpages work, it is often a better option to reply and explain to them. Lots of people come to high profile talkpages to troll or spam, and removal of those is fine. With someone who might become a real editor we try to be a little more flexible with how strictly we enforce the rules. The Wordsmith^{Talk to me} 19:29, 29 March 2023 (UTC)

  Are you really arguing in defense of biting the newbie? Chalk me up as one agreeing that this could have been handled a good bit better, and that there is no frigging useful reason to keep on taking swings (tacit or otherwise) at the IP? Ravenswing 20:59, 29 March 2023 (UTC)

It is not entirely clear to me what part of WP:NOTFORUM the IP address here violated. Can someone point it out to me, and explain how the IP violated it? Shells-shells (talk) 19:24, 29 March 2023 (UTC)

Hi, I'm ScottishFinnishRadish, and it seems I've been mentioned a lot in this section. It's likely that I've removed more talk page posts per NOTAFORUM and SOAP than everyone in this section combined, and I would have let that stand, or if I had removed it and was reverted, starting a discussion explaining why I had removed it. In my view, NOTAFORUM/SOAP is a tool to cut down on talk page disruption, rather than a bludgeon to police talk pages. I've been seeing it used a lot more often lately, and in some pretty dubious situations, e.g. after another editor has replied, and I think editors should really think about whether it is disrupting the talk page, rather than a bit FORUMy, before invoking it. Also, the IP agreed with me, so they can't be all bad, right? ScottishFinnishRadish (talk) 20:06, 29 March 2023 (UTC)

Oh, also I don't think necroing an old section is disruptive on it's face. Plenty of talk pages are slow enough where a few passersby will comment in a section over a year or more and eventually it will demonstrate consensus. ScottishFinnishRadish (talk) 20:10, 29 March 2023 (UTC)

Hence why I said depending on how long ago the last reply was. To me if someone replies to a post that hasn't seen activity in 10 years that's not all that constructive to me (usually it should be archived by then but some pages probably aren't frequented enough for archiving to be useful). ― Blaze Wolf^Talk_{Blaze Wolf#6545} 20:16, 29 March 2023 (UTC)

Not constructive is different from disruptive. ScottishFinnishRadish (talk) 21:07, 29 March 2023 (UTC)

This thread is of the "tense" and "strong feelings" variety. 61.8.194.45 (talk) 05:19, 30 March 2023 (UTC)

• Comment Can we agree to close this thread with a general advisory that we should make an effort to be a little more lenient with IPs and new editors when they are attempting to contribute in good faith? The essay Wikipedia:Don't come down like a ton of bricks seems especially relevant here. Nobody has suggested any sanctions against either Blaze Wolf or IP47, so it seems like our time would be better spent elsewhere. The Wordsmith^{Talk to me} 14:04, 30 March 2023 (UTC)

  Agreed! The messages on their talk page can help them! Tails Wx 18:59, 30 March 2023 (UTC)

  No, please could we close it with a specific advisory directed at the relevant editors? This is a teachable moment.—S Marshall T/C 10:06, 2 April 2023 (UTC)

  I support either outcomes here. I myself would only remove talk page comments that are unambiguously vandalism or disruptive, and I don't see how the comments that were being removed here fit that category at all. AP 499D25 (talk) 03:36, 3 April 2023 (UTC)

  Y'all need to do it like this: User talk:CEvansMCO#Draft:Brett Stewart (musician) —Alalch E. 23:40, 2 April 2023 (UTC)

  4,138 bytes...sigh. Tails Wx 12:37, 3 April 2023 (UTC)

  Meet you in the middle at 3.3k? —Alalch E. 13:08, 3 April 2023 (UTC)

  Joking aside, I agree with S Marshall. Support closing with a balanced specific advisory. —Alalch E. 13:22, 3 April 2023 (UTC)

  How about something along these lines?

  In order to strengthen Wikipedia's third pillar, we advise Blaze Wolf to try to welcome newcomers to Wikipedia. In particular, he should revert only when necessary, refrain from edit warring in talk pages, and answer clarifying questions about his application of Wikipedia's policies. Additionally, we advise 47.232.88.19 (talk) to contribute more to discussion than a simple poll by suggesting specific changes based in Wikipedia policy and guidelines.

   — Freoh 02:35, 6 April 2023 (UTC)

  I used to welcome newcomers before I ever warned them. I stopped after I read something (I don't recall where) that said that welcoming didn't really do much to increase editor retention or something. So I stopped because I didn't see much of a point. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 02:45, 6 April 2023 (UTC)

  I personally don't think a specific targeted/logged warning is necessary here. Most of the issues mentioned above (edit warring etc) don't really seem relevant to this circumstance. Regarding the Welcoming issue, it may or may not increase editor retention in the statistics, but can be a nice thing to do and takes up very little time especially with Twinkle. I think if I could give you a message to take to heart, it would be that most people (and by extension, most new Wikipedians) are genuinely trying to do the right thing and be a good person. Wikipedia is a Byzantine nightmare of complex policies, guidelines and other conventions that us veterans know but aren't even written down anywhere. If you see somebody who is technically breaking the rules but looks like they might be doing it because they don't understand the rules and are doing the best they can, it can be more helpful to leave them a personalized note explaining what they did wrong and how to do it right than dropping templates or ALLCAPS links to our (very long and dense) policies and telling them to read it. It costs nothing to be nice. The Wordsmith^{Talk to me} 03:04, 6 April 2023 (UTC)

  Also, holy run-on-sentence, Batman! The Wordsmith^{Talk to me} 03:07, 6 April 2023 (UTC)

• Comment I mean, the microcosm of this entire issue is in the first couple comments. The entire purpose of the "Put a notice at the persons talk page when you start a section about them" rule is to make sure people are aware they are being discussed. It is mind-boggling to me that you would respond to a new editor doing their best to make you aware of this (while already being aware of the discussion!), with "ah ah ah, you didn't do it at the right page! resubmit the forms, please." We don't have rules for the sake of rules, we have them for REASONS. When an editor, especially an IP!, attempts to follow the spirit of the rule in good faith, if not the letter, continuing to argue with them about shows that you need to take a step back and re-evaluate why you're enforcing this rule. Is it for the benefit of the wiki and its users? Parabolist (talk) 07:56, 7 April 2023 (UTC)

  It is mind-boggling to me that you would respond to a new editor doing their best to make you aware of this (while already being aware of the discussion!), with "ah ah ah, you didn't do it at the right page! resubmit the forms, please." That's something that is required at ANI. If I hadn't said it someone else would've. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 13:54, 7 April 2023 (UTC)

  Well done proving the point. --JBL (talk) 17:15, 7 April 2023 (UTC)

  Multiple misunderstandings there. The rule exists and for a reason. The IP (mistakenly) assumed that making an effort was enough and that Blase Wolf was going "over the top" on strictness or wrong in saying that it wasn't met. Blase Wolf responded by more pointedly saying that they were correct and what needed to be done differently. I would have done the same, albeit probably with a bit more explanation. North8000 (talk) 17:42, 7 April 2023 (UTC)

  But in this case, it WAS enough. Blaze Wolf was already aware of the discussion. If the person is already aware by other means, why bother enforcing the rule in this case? What is the benefit, and if the only benefit is "well they'll know next time", that is absolutely not what happened here. Its not a huge deal, but the fact that Blaze Wolf above still doesn't understand the point I'm making is the problem here. We don't have rules simply to enforce them. An IP coming here to complain about what they perceive as byzantine rules being used against having their concern be immediately set aside because they didn't file their form in the right bin is a user who is not being served correctly by ANI. Parabolist (talk) 18:56, 7 April 2023 (UTC)

  You seem to be misunderstanding. I only noticed this happened because I have ANI on my watchlist. If not then I never would've know. We have rules for a reason. having their concern be immediately set aside because they didn't file their form in the right bin Their concern wasn't set aside. I never even said that just because they didn't notify me correctly their concern was invalid. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 20:09, 7 April 2023 (UTC)

  Also, just because it was enough in this case, doesn't mean the rule gets to be ignored. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 20:10, 7 April 2023 (UTC)

  Why not? That's my point. They even told you that they made a good faith effort to do so, just not on the right page! ANI is the last port of call for users having trouble. So many threads by new users start with a veteran user chiding them over this, all the while never actually addressing the topic of their problem. You saw it, you were aware of it, and you responded solely to chide them for reporting you incorrectly, and then arguing with them about it, never once addressing the issue at hand! There's a lot of reasons ANI has the reputation it does, but this one specifically is one everyone here can personally work to make better. Parabolist (talk) 20:21, 7 April 2023 (UTC)

  I wanted to make sure they understood why I was arguing with them. I didn't see a need for me to respond myself since from my point it seemed like a waste of time and energy arguing further. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 20:28, 7 April 2023 (UTC)

  If your attempts to be helpful to new and IP users include "I think you need to take some time and read things completely so you actually understand what you are being told." you should probably leave that to another user not in conflict with the filer. Think about their experience from their perspective. Would you have found your responses helpful or friendly? Parabolist (talk) 20:35, 7 April 2023 (UTC)

  BW, That's something that is required at ANI. If I hadn't said it someone else would've. Or possibly what a lot of people would have done with an editor with a handful of edits is say, "IP, you're required to notify the editor at their User talk:Blaze Wolf; I have done that for you." And from the discussion at their talk, they make a very good point. They made a single short comment agreeing with someone, so you just pelted them with policy? Valereee (talk) 20:41, 7 April 2023 (UTC)

  Do we want to close this or are we just going to continue telling me that my behavior was wrong? ― Blaze Wolf^Talk_{Blaze Wolf#6545} 21:38, 7 April 2023 (UTC)

  Well you could do the decent thing and say "Yes I admit my behavior was not ideal, I shouldn't have done that" instead of digging in and then this whole thread of conversation will go away. --JBL (talk) 22:36, 7 April 2023 (UTC)

  I'm not attempting to dig in. It honestly just feel like you guys just keep piling this on top of me. Yes I understand that I could've done better but you don't need to keep telling me that I should've told them something else. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 22:41, 7 April 2023 (UTC)

  Literally this is the first comment that you've made that acknowledges the validity of the complaints, and you didn't even make it to the end of the sentence without switching back to defensiveness. If you want people to take the view that you understand the complaints, recognize their validity, and will take them to heart, you should try saying things that indicate that (instead of being defensive and making excuses). [This advice is extremely generalizable -- e.g. my spouse and I have improved our marriage by adopting less defensive modes of communication in the face of perceived criticism, and we don't even edit an encyclopedia together.] --JBL (talk) 22:46, 7 April 2023 (UTC)

• Comment: I would like to add to this discussion that anyone who was involved in this incident who also has rollbacker rights should know better than to revert anything other than blatant vandalism on a talk page since they have been quizzed extensively in order to get those rights that they have read all the documentation regarding the use and privileges of those rights. I understand people can get swept up in a kind of "righteous indignation" about being over zealous with the rules, and make mistakes, as I myself have done, but we must acknowledge them, and learn from them, then grow from there. Huggums537 (talk) 21:28, 9 April 2023 (UTC) I would also like to add to my comment that just because someone did not "technically" use the rollback feature specifically to do a reversion does not exonerate them from anything since owning that right is proof that they have educated themselves on which reversions are considered best practices. So, whether they did or didn't use the tool is really just a technicality since there is evidence they should still know better either way. Huggums537 (talk) 19:06, 10 April 2023 (UTC)
• Comment: Another thing I would like to add is that it is very hard to assign any blame to Blaze Wolf for their actions when I found out that just recently a related ANI discussion was closed in their favor, which might give the false impression that these kinds of reversions are appropriate when they really are not. It should have been more clear the OP was actually blocked for being NOTHERE and a sock, and only two admins did that correctly in that discussion. Huggums537 (talk) 01:45, 11 April 2023 (UTC)

Ngăn chặn kiến (talk · contribs · deleted contribs · logs · filter log · block user · block log) is making numerous small, at best pointless, edits to many articles. Adding or removing a comma here or an apostrophe there, putting a full-stop after "see also", that sort of thing. No edit summaries. Does the behaviour ring any bells? THanks, DuncanHill (talk) 13:40, 11 April 2023 (UTC)

• I checked out the user's most recent three edits and didn't see anything of concern. If you have a specific example I'm confident several of us would check it out. Have you noticed any harm or policy violation, or just editing preference?--Paul McDonald (talk) 14:12, 11 April 2023 (UTC)

• Paulmcdonald Nearly all of their edits are simply incorrect and have had to be reverted. I have left them a final warning per WP:CIR. Black Kite (talk) 18:20, 11 April 2023 (UTC)

  I trust you. I only looked at three and gave up when I didn't see a specific example.--Paul McDonald (talk) 21:02, 11 April 2023 (UTC)

user:Northheavensky is reluctant to communicate and accept the agreed consensus on the talk page.

Talk:Songket#Bias User:Ckfasdf and I had actively discussed on the talk page for days and we had agreed on NPOV for the page Songket

user:Northheavensky is reluctant to communicate or accept the agreed consensus by making the below edits:

- no reason for reverting was provided

- the user wrote "Stable version, radical changes without consensus, the source is still in doubt". This is a baseless accusation and has been discussed. We literally cite the same existing source, in the same book and had already reached a consensus on the talk page.

- no reason for reverting was provided

- no reason for reverting was provided, except "restore". MrCattttt (talk) 22:19, 11 April 2023 (UTC)

• The discussion on the talk page doesn't cover all the edits that the OP has made, though. Meanwhile, some of the edits by the OP (some of whose previous edits have been revdeleted by User:Diannaa for COPYVIO issues) appear to be in some cases poorly sourced and many of them remove sourced material. Therefore, I have fully protected the article in its current state (which appears to be pretty close to the long-standing version before the OP started editing it) so that an actual consensus can arise. Black Kite (talk) 22:57, 11 April 2023 (UTC)