User talk:Gnomingstuff

May the bells of Christmas ring for freedom!^[1]

Carol of the Bells (Instrumental)

Instrumental arrangement of "Carol of the Bells"

---

Problems playing this file? See media help.

May peace be upon us.

And have a happy and prosperous New Year.

7&6=thirteen (☎) 15:57, 21 December 2025 (UTC)

As you were the one who put it there, I would like to check with you before removing the template in question. I restored an older version of the "works" section, renamed it to "views" and added a few minor tweaks to phrasing that I believe improve sentence flow. Comrade GC (talk) 22:44, 21 December 2025 (UTC)

It's a judgment call on your part. Gnomingstuff (talk) 01:19, 22 December 2025 (UTC)

who is the AI editor there? i think you can be blunt in stating it on the articles directly as it is most conducive for the original goal of having the template removed again. EnTerbury (talk) 23:47, 21 December 2025 (UTC)

The goal is to fix the article, not to remove the template. I don't know why everyone hyperfixates on the template and not the actual core issue. Gnomingstuff (talk) 05:25, 22 December 2025 (UTC)

perhaps because it is easier to fix the article knowing exactly which parts constitute the core issue without having to check every single sentence and every single source, which seems rather disproportionate. if you are afraid that there may have been other LLM editors, you could slap it on every article. if you have reason to suspect it, name it. EnTerbury (talk) 07:38, 22 December 2025 (UTC)

to be clear, you are right of course that the goal is to fix the article. it is what i meant with removing the template. EnTerbury (talk) 07:52, 22 December 2025 (UTC)

There are guidelines on Template:AI-generated as to how to handle it, as well as Wikipedia:WikiProject_AI_Cleanup/Guide#Cleaning_up, but the gist is that every AI-generated edit needs to be checked line by line/fact by fact. It's a lot of work, but in the sense that it's work that should have been done before the text was inserted into the article, and in many cases of AI text was not. The edits in this case are by EEJCC (see their talk page, plus some buggy reference cruft left in edits like this.) Gnomingstuff (talk) 04:15, 23 December 2025 (UTC)

i guess these links should be in the template instead of the generic template cleanup article EnTerbury (talk) 05:38, 23 December 2025 (UTC)

Merry Christmas and a Prosperous 2026!

Hello Gnomingstuff, may you be surrounded by peace, success and happiness on this seasonal occasion. Spread the WikiLove by wishing another user a Merry Christmas and a Happy New Year, whether it be someone you have had disagreements with in the past, a good friend, or just some random person. Sending you heartfelt and warm greetings for Christmas and New Year 2026. Happy editing, Abishe (talk) 09:13, 25 December 2025 (UTC) Spread the love by adding {{subst:Seasonal Greetings}} to other user talk pages.

Abishe (talk) 09:13, 25 December 2025 (UTC)

I'm asking you because you seem like pretty much the most knowledgeable person on these matters. In Draft:Michele Friend, do you think the |degree=PhD Thesis could be seen as an AI sign in its own right (and could therefore be listed somewhere on WP:AITELLS), or is it something that happens roughly as often with humans who read the template docs but don't use preview? lp0 on fire () 21:47, 25 December 2025 (UTC)

(I'm tagging the article in question for G15, so here^[1] is the reference in question.

lp0 on fire () 22:33, 25 December 2025 (UTC)

Hi, sorry I didn't get back to you earlier, it's the holidays. I don't think that it's really indicative one way or the other. Gnomingstuff (talk) 04:09, 26 December 2025 (UTC)

Thanks. No worries about response time; I was just asking out of curiosity. Happy holidays :) lp0 on fire () 10:24, 26 December 2025 (UTC)

I think that the frequency and similarity of these mistakes, from a single editor, can be a tell. But you have to exercise judgement and take other tells into consideration as well.

Something that I have noticed is that LLM generated drafts tend to be re-"written" semi frequently, where the user is asking the LLM for a new version of the article ("please refine this to.."), and then pasting the new version in place of the older one. That often shows up as edits that touch many areas of the same draft in rapid succession of each other. These invalid parameters can come and go in those types of edits. --Gurkubondinn (talk) 08:26, 29 December 2025 (UTC)

Hi. You suggested I used AI for my edits of the Marx Brothers article. I spent a fair bit of time researching my changes, and stand by what I added, so I would really rather the notice not be there - at least, not because of anything I wrote. What can I do to convince you that the notice doesn't belong there? Are there specific edits you would like to see?

One thought: I feel like some of the text you consider to be an AI tell might be close to violations of WP:NPOV (e.g., the changes to the legacy section you pointed out). If I do a pass over the article to remove the POV material, would you be comfortable with my removing the template? Jm307 (talk) 10:50, 26 December 2025 (UTC)

I mean the first piece of missing information here is whether you used AI or not. Gnomingstuff (talk) 16:27, 26 December 2025 (UTC)

I drafted the changes and - for the larger ones - used AI to check my grammar and offer suggestions on how to improve what I had written. I took some of the suggestions, and ignored others. I definitely ignored any suggestions that were factually wrong or otherwise not in keeping with my then-understanding of what makes a good Wikipedia article.

It wasn't too different from using a spelling or grammar checker. The text is mine, other than to the extent that the LLM suggested some word choices and different phrasing.

The results are hallucination-free, although I can see how certain phrases that it suggested and I deliberately incorporated would have triggered your search.

I spent time with a lot of reference material on the Brothers to make the edits I made. I have more edits planned. The template you added suggests to the reader that the information on the page is unreliable, but I made sure that those changes were in keeping with reliable sources.

That said, since I made those edits, I got my first experience ushering an article through the GA process, and recognize text that the LLM suggested that does not belong on Wikipedia. I can make a pass through the article and remove that text, but I want to make sure that that's going to be enough. Jm307 (talk) 20:55, 26 December 2025 (UTC)

I just noticed WP:NEWLLM, which says that LLMs should not be used to generate new Wikipedia articles from scratch. To be clear, in this case, I did not generate any of these changes from scratch with an LLM. Jm307 (talk) 21:02, 26 December 2025 (UTC)

I have now done a very manual, AI-free audit of the page to remove the offending AI-generated material. I double checked what I added, and everything is factually correct. There remains much material on the page that is not supported by sources, but there is no reason to think it was added by AI.

Do you have objections to my removing the tag? Jm307 (talk) 00:21, 28 December 2025 (UTC)

Yes, I have an objection. The tag is factually accurate. If AI was used in an article, readers deserve to know that. If you don't like readers knowing you used AI, then don't use AI. Gnomingstuff (talk) 16:36, 28 December 2025 (UTC)

I know you are getting a lot of very strong pushback, and I don't want to seem disrespectful of what you are trying to do. Keeping wikipedia free of hallucinations is very important. However, there is a clear difference between using AI as a tool to improve the clarity and structure of something that you've already written, and using it to come up with new content.

If the template only said "this may have used AI", I would have less of a concern with your argument, but it also says "It may include hallucinated information, copyright violations, claims not verified in cited sources, original research, or fictitious references. Any such material should be removed, and content with an unencyclopedic tone should be rewritten." I have carefully audited my use of AI to ensure it included none of these things, and rewritten or removed the content that had an unencyclopedic tone. The template leads the reader into thinking that the contents of the article may be inaccurate because of the use of AI. As it is, it is no more inaccurate than it would have been if I had not used AI. Jm307 (talk) 23:07, 28 December 2025 (UTC)

Or, to put it another way, the template says "It may include hallucinated information, copyright violations, claims not verified in cited sources, original research, or fictitious references. Any such material should be removed, and content with an unencyclopedic tone should be rewritten." I have removed such material and rewritten the content with an unencyclopedic tone. What more is there to do? Jm307 (talk) 23:12, 28 December 2025 (UTC)

Gnomingstuff didn't come up with the template language that so is not something you should bring up with them. Given that you used AI to assist with your edits, I think the application of the tag is quite appropriate. There is overwhelming evidence that most people who use AI to assist with their article-space edits do not sufficiently review the LLM-generated content prior to adding it to the article. My suggestion would be to ask a 3rd party for an independent review of some of your LLM-assisted edits, checking them for issues with WP:V, WP:NPOV, and WP:SYNTH. If they don't find any, then the tag can be removed and you will have support in future conversations around your LLM use. NicheSports (talk) 23:14, 28 December 2025 (UTC)

I'm happy to have an independent audit. That said, most people will not have access to the sources I used (I have a decent sized library on the topic). While I think my last edit should help it pass the WP:NPOV and WP:SYNTH smell tests (and am happy to make further edits to improve that), I don't think the average Wikipedian will be able to check my sources. Jm307 (talk) 23:46, 28 December 2025 (UTC)

Also, I'm not sure whom / where to ask for an independent audit. Guidance would be appreciated. Jm307 (talk) 23:47, 28 December 2025 (UTC)

I typically don't do this but will in this case. As long as you are fine with me doing it next weekend. I may need some help in reviewing sources - I think it can sometimes be appropriate to share them privately (this is done in GA review processes where sources are hard to find). Ok if you ping me at my talk page? NicheSports (talk) 00:37, 29 December 2025 (UTC)

That's kind.

I am traveling through Jan 6, so I won't be able to help with sources until then. Most of them are books, and I cited quite a lot of individual pages, so hopefully I won't have to do a bunch of photocopying. I know at least one of the books is available on IA, but I used my copy, which is a different edition, so the pages are off.

One challenge will be that the material I contributed is not up to GA quality (although is closer than what was there), so won't stand up to a GA-quality review. I focused on:

• Edits to the lead.
• Bulking up the section on their lives through the movie Monkey Business. This is where I added the most new information and references. I did this before going through my first GA review; I learned a lot from that process, and suspect that the material I added won't stand up to that kind of scrutiny.
• Trying to turn the Influence section into less of a trivia dumping ground. This is going to be a bit annoying to review, because I continued to use a lot of the uncited information that was already in there without adding citations. If it is wrong, it is not because of AI. That said, I continued to use a lot of the information and citations that were already there, so that might checking the refs easier (if it was there before my changes, it's definitely not something I added with AI).
• Organizing the Theater and Filmography sections. I did not use AI for this, so you can skip it if you like.

When a reference was involved, I used AI to suggest better wording and organizational structure, but did not copy / paste what it generated, specifically because the LLM couldn't deal with the ref correctly. (Most of what sounded like AI-generated content was connecting sentences and paragraphs that I have now deleted.)

Quick diff from when I did most of my edits:

https://en.wikipedia.org/w/index.php?title=Marx_Brothers&diff=1289554649&oldid=1284415314 Jm307 (talk) 23:22, 29 December 2025 (UTC)

The page https://en.wikipedia.org/wiki/Nasandratrony was deleted in bulk by Gnomingstuff with the rationale “AI slop.” Regardless of the means of drafting, the removed content consisted of verifiable, factual statements supported by reliable, independent sources that were explicitly cited. Wikipedia policy evaluates contributions based on verifiability, sourcing, and encyclopedic value, not on whether AI tools were used in the writing process. As the material met core content policies, its wholesale removal on the basis cited was not consistent with those standards, and the reverted text should be assessed on its merits rather than on the method of composition. Oroboros74 (talk) 16:00, 26 December 2025 (UTC)

I was the one who reverted you this time, like I said on the talk page, AI content is not desirable here, even as talk page messages. Plasticwonder (Cat got your tongue?) 16:04, 26 December 2025 (UTC)

I didn't delete the page. The page still exists. Only administrators can delete pages. Gnomingstuff (talk) 16:26, 26 December 2025 (UTC)

You've deleted entire sections - facts , statistics, references. You've deleted close to 22k words. That is not positively contributing.

I would like to suggest you go through each paragraph and make edits if you feel, but you should not delete entire sections just becasue they may have been drafted in part using AI. If you actually toook the time you'd see that everything has been cited and referenced, even to the page level.

AI was used to rephrase, but not for the facts. And if you don't trust this, feel free to fact-check each piece of information (I welcome it!) and document WHY/WHAT needs to be edited. Oroboros74 (talk) 16:34, 26 December 2025 (UTC)

From Wikipedia's five pillars:

All editors freely license their work to the public, and no editor owns an article – any contributions can and may be mercilessly edited.

If you don't like the principles of Wikipedia, then Wikipedia may not be the best place for your AI writing. Gnomingstuff (talk) 16:35, 26 December 2025 (UTC)

Do you think that throwing out the baby with the bathwater is the right solution?

I welcome any edits to the text you might have, as long as you're contributing positively. But deleting 22k words of documented facts is just wrong.

I challenge you to find one fact that is not fact checked or that is false. Oroboros74 (talk) 16:39, 26 December 2025 (UTC)

Yes, in the case of AI-generated text. This is a Wikipedia guideline.

Please stop spamming my talk page. Gnomingstuff (talk) 16:40, 26 December 2025 (UTC)

The text was not AI-generated. There are no hallucinated facts. AI was used for an initial revision, and you can look at the revision history of the page and see edits even on a word level.

Please stop deleting 22k words from the wikipedia page I've been actively maintaining over the past weeks.

If you find something that is not fact checked or false, I welcome the edit, but stop this war on this poor African village's wikipedia page. Oroboros74 (talk) 16:59, 26 December 2025 (UTC)

For the love of god stop spamming my talk page, stop pinging me just to whine and whine and whine, leave me the fuck alone Gnomingstuff (talk) 17:00, 26 December 2025 (UTC)

I suggest you drop the stick, Oroboros74. Take this as a lesson to stop pestering those who disagree with you. Your assertion that you did not use AI is not true by your own admission, take a step back and do better. And I second Gnomingstuff in imploring you to stop whining. Plasticwonder (Cat got your tongue?) 23:05, 26 December 2025 (UTC)

"Regardless of the means of drafting" - Actually, how the page was written is important. AI very, very frequently completely makes stuff up. --not-cheesewhisk3rs ≽^•⩊•^≼ ∫ (pester) 16:30, 26 December 2025 (UTC)

Please read carefully all the edit notes and talks, and make no mistake. All facts have been sourced and referenced. And yes drafted - which means that the final version was written using human judgement not artificial intelligence. Oroboros74 (talk) 16:36, 26 December 2025 (UTC)

Did you use AI to draft the article? --not-cheesewhisk3rs ≽^•⩊•^≼ ∫ (pester) 19:28, 26 December 2025 (UTC)

I suspect that Oroboros74 used AI, per the almost mechanical talk page messages here. I do not mind as certain users may not have a perfect mastery of english, but it appears that is not the case for Oroboros74, who claims to not use AI. Plasticwonder (Cat got your tongue?) 19:43, 26 December 2025 (UTC)

Could this discussion please take place somewhere other than my talk page? Gnomingstuff (talk) 05:30, 27 December 2025 (UTC)

While your contributions towards cleaning up AI-generated content on Wikipedia are appreciated, please be more careful before tagging entire articles as AI-generated. Your flagging of Mean (song), Liyue (Genshin Impact), and Seven (1995 film) are not acceptable and almost come off as personal attacks against the editors involved. Especially Seven and Mean (the latter of which you claimed was entirely AI-generated but did not provide suitable evidence for). It's also made worse when you A) don't bring it to the talk page of the responsible editors, and B) don't discuss the matter further on the talk page at all. You drove-by these articles, tagged them as AI using flimsy, broad justifications (mostly based entirely on simple word choice), and then did not discuss the matter any further. Please communicate.

Many times, the articles or edits you clean up as "AI-generated" do, in-fact, seem AI-generated. Your edits are mostly constructive in this regard. But this is not the approach. Please be more careful when tagging entire articles as AI-generated. This could be by either asking the editors involved first if they used AI or not, or pointing out more convincing signs that an article is AI-generated, or just being willing to discuss the matter further on the article talk page itself. Thank you. λ NegativeMP1 19:26, 28 December 2025 (UTC)

There is currently a discussion at Wikipedia:Administrators' noticeboard/Incidents regarding your efforts to clean up AI slop on Wikipedia. The thread is Gnomingstuff and AI cleanup. Gommeh 📖   🎮 19:37, 28 December 2025 (UTC)

Hi...i have been one of the editors on Science and technology in Jamaica - Wikipedia. The drafting/generating of the original content/text is done by human. Is there an issue with using AI simply to help improve the wording and phraseology of the human generated content/text? My edits are properly sourced and referenced (and not hallucinated). Feel free to make any edits or changes you think are needed...or perhaps you could point out the bits you think need changing/removing/or need better references? Many thanks. Koppite1 (talk) 00:04, 30 December 2025 (UTC)

Hi,

As i've had no feedback or response, i'm inclined to remove the AI tag that you put on the article. I'll reiterate that the original content wasn't drafted by AI. Also everything is sourced, with links. Many thanks Koppite1 (talk) 10:30, 22 January 2026 (UTC)

An automated process has detected that when you recently edited WeChat Pay, you added a link pointing to the disambiguation page JCB Co., Ltd..

(Opt-out instructions.) --DPL bot (talk) 07:56, 2 January 2026 (UTC)

Hello, Gnomingstuff. Please check your email; you've got mail!
Message added 23:49, 4 January 2026 (UTC). It may take a few minutes from the time the email is sent for it to show up in your inbox. You can remove this notice at any time by removing the {{You've got mail}} or {{ygm}} template.

LR.127 (talk) 23:49, 4 January 2026 (UTC)

I’m a bit concerned that your recent messages at the Brittney Spears talk page and here might fall under inappropriate canvassing as described at WP:CANVASS – in particular, the way specific editors were invited and the framing of the discussion.

Could you please review that guideline and consider using more neutral, broadly targeted notices (for example, at a relevant noticeboard or WikiProject) instead of contacting selected editors in this way? This helps keep discussion balanced and avoids any appearance of vote-stacking.

If you think I’ve misunderstood the situation, I’m happy to discuss it.

This case is particularly troubling since Gnomingstuff actually followed through and wrote on the Brittney Spears talk page that he got pinged to enter that particular discussion at that time. BassiStone (talk) 11:54, 6 January 2026 (UTC)

 You are invited to join the discussion at Talk:Britney Spears § Shall we remove the 2019-2021 personal life section?. sjones23 (talk - contributions) 19:38, 5 January 2026 (UTC)

Thanks for the reminder on my talk page. As a longtime Wikipedian, I'm only trying to help out the best I can.

About that Britney Spears situation, I became very suspicious of the AI/LLM involved and removed all of BassiStone's edits trying to help improve it, but since my edits were reverted, I asked around to see what I can do to resolve the situation. If a disruptive user persists they're not trying to cause any problems, I would try asking somewhere to get another opinion. However, when I brought up the AI/LLM situation, they posted hostile comments on your talk page and on my talk page, so I removed them.

All the best for the new year, sjones23 (talk - contributions) 23:40, 6 January 2026 (UTC)

No need to apologize. This may not have been the best thread to be pinged into but I don't think you had any way of knowing that. Likewise with the talk page stuff. Gnomingstuff (talk) 00:20, 7 January 2026 (UTC)

I see. After all, we're all volunteers and people too. Even editors who have been around for 19 years such as myself make mistakes from time to time. sjones23 (talk - contributions) 04:55, 9 January 2026 (UTC)

Hello Gnomingstuff! I understand your good intentions and effort in detecting AI written material in the Wiki. However, due to your edit in the Eastern esotericism article, I ask you to take your attention to these pertinent questions before making your judgments: first of all, the public use of LLMs and their popularization only started after November 2022. The article in question was first created in Portuguese in August 2022, and it was later translated in the beginning of 2023 to English by another user. Also, most written text may resemble AI due to a partial use of machine translation by people who don't speak English (which may be involved in translated articles, such as this one); or it may just be a matter of coincidence of mannerisms, since AI replicates (at least partially) real world writing. Also, discuss in the article pages with the community first, otherwise you will be unilaterally tagging and creating controversial and even prejudicial edits. Thank you, best regards! Bafuncius (talk) 15:25, 11 January 2026 (UTC)

I specifically linked to the edit in question in the tag: this. This edit was made on 14:10, 23 October 2024. For those keeping track at home, 2024 is later than November 2022. Please take the minimal amount of time that it takes to read the template and look at the relevant datestamps before deciding to lecture me on my talk page. Gnomingstuff (talk) 19:22, 11 January 2026 (UTC)

The tag is still misleading. It is unproportional for only one edit: it mischaracterizes as if a large part of the article was written by AI (which it wasn't, as I've stated). If the only problem is the lead, you only need to revert to the original one - so I'll be restoring it and removing the tag. Bafuncius (talk) 19:29, 11 January 2026 (UTC)

Feel free to restore the old lead. I did not revert it to an older version as there were many intervening edits since October 2024. When there are intervening edits, unless they are minor bot tweaks, I err on the side of not overwriting other people's content.

However, in this case, there was no other choice for where to put the tag, as the section= parameter only applies to specific sections, i.e., subheaders; there is no "Lead" header, it's just the lead. The template merely states "This article may contain text from a large language model." This does not mean THE ENTIRE ARTICLE, any more than "this product may contain nuts" means that the whole product is nuts. Any interpretation otherwise is misreading it. Gnomingstuff (talk) 19:32, 11 January 2026 (UTC)

I understand, however imagine that in the article there was only a single AI-written phrase or a shibboleth word created by AI: it would be misleading to feature the tag as a caution notice just because of it, and it would discredit the whole article. So, in lead sections, it is better to just restore or remove the AI content, instead of adding that scary tag. Bafuncius (talk) 19:41, 11 January 2026 (UTC)

I understand that; if it's only a minor change I usually just fix or revert it, or alternatively just don't tag if the change is minor/ok (some copyedits fall into this category). I tag it when it's not a minor change or when there have been a lot of other people making changes to the text since then so as to not overwrite their work. Gnomingstuff (talk) 23:39, 12 January 2026 (UTC)

Was wondering what tool you're using so I could use the same to see what results I get vs the free Al tool I am using. Best regards. Moxy🍁 23:30, 12 January 2026 (UTC)

Hi - I don't use automated detectors for the most part. I mostly look for the things identified in WP:AISIGNS and cross-reference them by the date of the edit -- there are some things there more characteristic of 2023-2024 LLMs, some things more characteristic of 2024-2025 LLMs, etc. Gnomingstuff (talk) 23:33, 12 January 2026 (UTC)

sounds reasonable..... You seem to be doing a great job. I will review the article again. Moxy🍁 23:38, 12 January 2026 (UTC)

No worries. Which article is this? Gnomingstuff (talk) 23:39, 12 January 2026 (UTC)

Tourism in Canada ...I did use ai summarizer for the promotional content that was already there (it was tagged for promotional tone)..... with the setting of removing promotional tone. I guess it came out worse than I thought. Have gone through quite a bit.... I'm getting a hit for the summary for the statistics Canada stuff... But that was copied directly from statistics of Canada.Moxy🍁 23:45, 12 January 2026 (UTC)

Ah thanks, and thanks for the explanation. LLMs aren't great in general at removing promotional tone even while copyediting since their "neutral setting" seems to be promotional, so a lot of times the edit will essentially be rearranging deckchairs. Gnomingstuff (talk) 01:31, 13 January 2026 (UTC)

yeah I'm coming to realize that they simply just don't work the way you think.... No matter how much oversight there is. Moxy🍁 01:58, 13 January 2026 (UTC)

Hello, I noticed you've left the suspected AI edit tag on the article University of Tokyo, and that one of my edits from a few years ago was pointed out as an example. All of the research used in my edits was done by myself and I did not make any claims that were not mentioned in the cited sources. However, I did use LLMs to make sure my edits were grammatical and sounded natural, as I did not have a comprehensive understanding of Wikipedia editing standards at the time and English is not my first language. I've just had a look at what I added back then, and it sure included unnecessary promotional words and expressions that were most likely recommended by LLMs at the time.

I shall work on the article to remove unsuitable wordings as well to add further verifications, so I would appreciate it if you could remove the tag when you think enough clean-up is made. VersedVoyager67 (talk) 02:29, 13 January 2026 (UTC)

Regarding this edit of yours:

1. Did you mean 2024 instead of 2023?
2. Would it be a good idea to revert to the pre-AI edit version then try to restore some of the intervening major edits? Some, not all, because going through all of them would take far too long.

Un assiolo (talk) 14:03, 16 January 2026 (UTC)

Hi - apologies, yes, I did mean 2024. (They're the ones with "in collaboration with ChatGPT" in the edit summary.) Up to you whether you want to revert, I usually err on the side of not reverting if there are a lot of intervening edits especially if it is a subject I am less familiar with.) Gnomingstuff (talk) 23:47, 19 January 2026 (UTC)

Hello, Gnomingstuff. I have noticed that you have been tagging articles with AI tags, including the page for Borderline personality disorder. The edit you tagged does not explain, either in the tag or on the talk page, what you believe justifies this tag, other than to link to thw advice page WP:AISIGNS. I went through each and every reference added appears legitimate. The user in question also has a long and positive contribution history and uses 2017 wikitext editor.

Understand, please, that I am also opposed to any AI-generated content on Wikipedia. However, whatever context that was used to justify this appears to be missing and it makes it difficult to follow your reasoning. I believe your intentions are good but I, personally, cannot see what in the edit you linked has raised your ire. Given some of the feedback you have received above in your talk page, I want to suggest very politely that you might be a little heavy-handed with your approach. The Pony Toast 🍞 (Talk) 03:30, 18 January 2026 (UTC)

As I exceedingly clearly and unambiguously stated in the template, the edits in question are the ~2024 edits such as this. As I also stated clearly and unambiguously in the template, they display marked signs of AI use, which I have expanded upon on the talk page, which is a more appropriate place for this discussion to happen than my personal talk page.

The user's edits before this are not relevant here because by definition anyone using AI tools did not use them before those tools existed. Nor is their choice of editor, you can paste AI content into anything. Gnomingstuff (talk) 23:41, 19 January 2026 (UTC)

Hi there,

I'm leaving this message because you contributed to the recent RfC regarding the inclusion of airport destination lists. As promised, now that that RfC has closed, I've initiated a further discussion about the sourcing standards to be applied to these lists.

If you wish to contribute to the discussion, please do so at Wikipedia:Village pump (policy)#Airport destination lists - sourcing requirements.

Cheers! Danners430 _{tweaks made} 15:17, 19 January 2026 (UTC)

I used AI on Teletriumph. I did between 8 to a dozen articles using AI around that time. On my userpage, I have also recently posted my 2026 Anti-AI Pledge. Please see: User:Guylaen#2026_Anti-AI_Pledge.

I have some work to rewrite the articles as human – but trust me that none of that article was hallucinated. Guylaen (talk) 12:11, 20 January 2026 (UTC)

But I should also say that "content expansions" should not be a justification. I still write a ton even completely human, as you can see in any of my recent edits. I just think most articles are too small anyways. 6,000 words is a stub for some subjects. Guylaen (talk) 12:19, 20 January 2026 (UTC)

Hi, thanks for the note. When I said "content expansion" that was just to identify which edits the tag referred to -- normally I link directly to the diff, but in this case there were a lot of them. It wasn't meant to imply that any article expansion is AI, that would be silly as you noted. Gnomingstuff (talk) 00:13, 22 January 2026 (UTC)

In August 2025 you made a suggestion to The Signpost submissions desk at Wikipedia:Wikipedia_Signpost/Newsroom/Suggestions#Suggestion_by_Gnomingstuff_(2025-08-18).

I referenced that when I invited you and other members of a WikiProject to submit a story to Signpost - Wikipedia_talk:WikiProject_AI_Cleanup#LLM-generated_article_published_in_The_Signpost

I am here to offer you a personal apology. I also apologize for this apology, as it is low quality. Despite me being one of just a few Signpost editors, I do not stay on top of reader submissions or suggestions, and usually never read them, as I did not read yours. As you noted, there is interpersonal conflict in Signpost editorial spaces about AI/LLM policy. I see that, and I would prefer in Signpost and elsewhere in Wikipedia more civility enforcement. Stricter enforcement could make things better, or it could make things worse, I am not sure, but in any case I understand why you would want your distance.

Even as I am apologizing here, I would also like to ask for a favor for you. Even if writing for Signpost is not a fit for you, I would like for you to have some kind of tolerance for the publication and refer anyone else to submit a story at any time in the coming months and years, if you see the opportunity. The Signpost is far overstretched, and frequently not of the quality that I want it to be, but it is a 20-year old newspaper that is intended to be by and for the people. If for whatever reason it is lost, then I do not think anyone would ever establish another one, and right now I think we are better off keeping it.

I respect your suggestion and I myself will continue to try to recruit a writer for the story you described. Thanks. Bluerasberry (talk) 15:34, 20 January 2026 (UTC)

Hi - there's no need to apologize, people miss things all the time. Thanks though. Gnomingstuff (talk) 23:16, 21 January 2026 (UTC)

An automated process has detected that when you recently edited Chikhli, Maharashtra, a link pointing to the disambiguation page Special was added.

(Opt-out instructions.) --DPL bot (talk) 07:53, 22 January 2026 (UTC)

Hi! Thanks for your interest in the article Dutch Maid Bakery. I added a section to its talk page to discuss "AI slop" issues, since I think I might have addressed the most of it, but would very much appreciate your input there to further improve the article. -- ke4roh (talk) 20:10, 23 January 2026 (UTC)

Hi! I noticed that you added the {{AI-generated}} tag to economic history of Mexico. I checked the page's history and it seems that almost none of it was written when LLMs could write text as sophisticated as what was in the article. This seems to be an obvious false positive. Thanks, Kora ^^ (she/her) ^{say hi!}/_{what I've done} 03:31, 26 January 2026 (UTC)

Hi, thanks for the note. I added the tag in August so I don't remember the rationale here. Gnomingstuff (talk) 04:13, 26 January 2026 (UTC)

(talk page stalker) See Talk:Economic_history_of_Mexico#c-Gnomingstuff-20250813182900-Possible_unreviewed_AI_use. * Pppery * _{it has begun...} 19:08, 26 January 2026 (UTC)

Thank you! Sorry I didn't see that. That seemed not very major/not worthy of putting a whole tag for when I was searching through the page history. I have tried to fix it now, it's not as peacocky as it was before and hopefully still as useful. Kora ^^ (she/her) ^{say hi!}/_{what I've done} 01:30, 27 January 2026 (UTC)

Thank you for your contributions to Wikipedia. Regarding your recent edits to Zsadány when you modified the page, you introduced unknown parameters. Just because you specify |some_param=some_variable does not always mean that variable will display. The |some_param= must be defined in the template. You can look at the documentation for the template you are using but it is also helpful to use the preview button before you save your edit; this helps you find any errors you have made and ensure that the values you have added are displaying correctly. Below the edit box is a Show preview button. Pressing this will show you what the page will look like without actually saving it. It is strongly recommended that you use this before saving. Note I have likely fixed the error by now so check the history of the page to see how it was fixed. If you have any questions, contact the help desk for assistance. Thank you. Zackmann (^{Talk to me}/_{What I been doing}) 03:05, 30 January 2026 (UTC)

I was just looking into the user you posted about at AINB, came across them while filing about an unrelated Nepal-based sock. Small world. Hope you're doing well, I'll be more active again starting next weekend. NicheSports (talk) 19:41, 6 February 2026 (UTC)

weird! No need to apologize for activity, life is more important after all. Gnomingstuff (talk) 20:13, 6 February 2026 (UTC)

Hello, for the article William W. Destler, can you please expand on your AI generated flag edit . I'm not sure I see the Claude connection with the main page diff link you provided, and an IP editor just removed the flag on Destler's page. Still trying to learn how to detect these things. Cheers, --Engineerchange (talk) 18:52, 7 February 2026 (UTC)

Kind of an unusual case, in this case the main reason we know about it is that someone bragged on Reddit about it, and the temporary account affiliated with that user has a lot of similar activity to other temporary accounts. The link might have been a typo on my part but this is the topic that links to the details: Wikipedia_talk:WikiProject_AI_Cleanup/Archive_5#h-LLM_powered_whitewashing-20260107143400 Gnomingstuff (talk) 04:21, 8 February 2026 (UTC)

Ah, okay, this makes sense. Thank you! --Engineerchange (talk) 05:24, 8 February 2026 (UTC)

Hello, I'm Qwerfjkl (bot). I have automatically detected that this edit performed by you, on the page Mental health provisions in Title V of the No Child Left Behind Act of 2001, may have introduced referencing errors. They are as follows:

• A generic name error. References show this error when author or editor name parameters use place-holder names. Please edit the article to include the source's actual author or editor name. (Fix | Ask for help)

Please check this page and fix the errors highlighted. If you think this is a false positive, you can report it to my operator. Thanks, Qwerfjkl (bot) (talk) 10:01, 12 February 2026 (UTC)

Regarding this edit: note by changing the prefix string of characters from * to ::, the wikitext parser closes, from right to left, a first-level bulleted list, and then opens, from left to right, two nested unbulleted lists. This causes screen readers to make a list end announcement and two list start announcements, versus just one list start announcement if you had used *:. In accordance with Wikipedia:Manual of Style/Accessibility § Lists and Help:Talk pages § Indentation, could you consider following the guidance to preserve the prefix string of the comment to which you are replying, adding the list character of your choice? (Using : for your response is less likely to result in others resetting the nesting hierarchy in subsequent responses.) Following this guidance would avoid extra pairs of list start/end announcements. You can also see User:Isaacl/On wikitext list markup for more details. Thanks for your consideration! isaacl (talk) 04:59, 13 February 2026 (UTC)

Thanks for the note, will keep in mind -- mostly I'm trying to handle old unsigned comments where it's confusing what text belongs to whom. Gnomingstuff (talk) 05:06, 13 February 2026 (UTC)

In this case, it's a comment you added, not an unsigned comment. Thanks for your co-operation. isaacl (talk) 05:32, 13 February 2026 (UTC)

Hello Gnomingstuff, I noted that you put an "AI-generated" template in the article Caritas Hungary. I have very much written this article myself, and it took quite a bit of research. I believe that everything stated in the article has a source - however it is true that I probably used AI to proofread the grammar and syntax of some parts of the article as English is not my first language. If I didn't do it with AI, then I might have used a translating tool for sentences that I wrote in German. I don't really remember as this was in 2024.

I am wondering what I can do to prove that the article itself is not written by AI and doesn't contain hallucinations or plagiarism? Should I adapt the language? Thanks for keeping the quality of articles high, at least now I know not to use AI for "improving" my language. --Bdx (talk) 14:18, 13 February 2026 (UTC)

If you used AI then the tag is factually accurate. Gnomingstuff (talk) 16:13, 13 February 2026 (UTC)

I do not agree, as the tag insinuates that the article contains "hallucinated information, copyright violations, claims not verified in cited sources, original research, or fictitious references", none of which is true. I'll rewrite the article. --Bdx (talk) 08:10, 16 February 2026 (UTC)

Hi there,

I took the liberty of cleaning up the talk page for WebMD and removed the unrelated threads as per WP:NOTFORUM and WP:TALKOFFTOPIC. Please let me know if i made a mistake. Thanks! Jratmain (talk) 20:07, 14 February 2026 (UTC)

With the text dating to at least 2011 can you explain why you think it is the product of an LLM?©Geni (talk) 20:47, 14 February 2026 (UTC)

This is the diff; it's from April 2025. I tagged this in August and it looks like it's been removed since, so the tag can also be. Gnomingstuff (talk) 23:29, 14 February 2026 (UTC)

Regarding this edit: could you consider a response that is more encouraging? Although I share your concern that the initial post seemed to describe a tautology (are Wikipedia users volunteers and people?) with specific presumptions, I think your followup doesn't encourage a collaborative answer. I appreciate your time in considering this matter. isaacl (talk) 17:34, 15 February 2026 (UTC)

The discussion was already blatant strawmanning intended to paint people who are concerned about AI text in a bad light, with zero concrete evidence, so right out of the gate there was not a presumption of collaboration.

Also, please don't spam my talk page and thus my pings with things that are better placed elsewhere. Gnomingstuff (talk) 17:43, 15 February 2026 (UTC)

It's common practice to have quiet words raising concerns with editors directly on their talk page. Thanks for your time and consideration. isaacl (talk) 18:09, 15 February 2026 (UTC)

And I have said multiple times, including on my userpage, that if there is any reason you don't have to to ping me -- which also generates an email I have to go delete from my inbox, and yes, I know that I can change settings, I just don't want to -- then don't.

I don't think this discussion needs to continue and I do not want another ping. Gnomingstuff (talk) 18:12, 15 February 2026 (UTC)

In the future, can you consider reverting LLM edits from editors like this vs. just leaving a {{AI-generated}} tag? Looking at it months later makes it much trickier to address the problematic LLM use. Ed ^{[talk] [OMT]} 05:21, 18 February 2026 (UTC)

I both consider and do this all the time. The problem is that it's very much a damned if you do, damned if you don't situation. If you revert people's edits, people complain. If you don't revert people's edits, other people complain. I have yet to find the action that will result in being (finally, blessedly) free of people complaining at me. Gnomingstuff (talk) 05:23, 18 February 2026 (UTC)

I think it's pretty clear that needing to revert articles back 6+ months is worse than doing it near the original problem, but in the end it's up to you. :-) Ed ^{[talk] [OMT]} 05:25, 18 February 2026 (UTC)

If the edits are very recent and it's possible to revert without overwriting other people's work, I do usually revert in those cases (and get yelled at for it, prolifically). But since I mostly work on finding older undetected AI edits, often there will have been hundreds or in some cases even thousands of edits made since then. Gnomingstuff (talk) 05:30, 18 February 2026 (UTC)

Gotcha. In the above case, it was just a couple months, and I am currently WP:TNTing them back to the pre-LLM state. It does quickly become too difficult to disentangle unchecked LLM content from legitimate, and to me it's not worth the risk to Wikipedia's reputation. Ed ^{[talk] [OMT]} 05:38, 18 February 2026 (UTC)

WikiProject Unreferenced articles | March 2026 Backlog Drive
There is a substantial backlog of unsourced articles on Wikipedia, and we need your help! The purpose of this drive is to add sources to these unsourced articles and make a meaningful impact. Barnstars will be awarded based on the number of articles cited. Remember to tag your edit summary with [[WP:MAR26]], both to advertise the event and tally the points later using Edit Summary Search. Interested in taking part? Sign up here.
You're receiving this message because you have subscribed to the mailing list. To opt-out of future mailings, please remove yourself here.

MediaWiki message delivery (talk) 11:13, 19 February 2026 (UTC)

That article you posted to AN/I was really, really good. Do you have a compendium of things like these somewhere I can peek around in? Or any links to more I might read? Likening to these sorts of examples as you did could prove a valuable tool in the future. Thank you!! Viva la horde, ~ GoatLordServant^(Talk) 17:24, 19 February 2026 (UTC)

Thanks -- unfortunately most of the stuff in the article took place on Twitter and as far as I can tell a lot of it's been deleted, and I don't know of anywhere it's archived. Gnomingstuff (talk) 17:42, 19 February 2026 (UTC)

Oh, that I understand! I mean, general high quality articles like that one. Not necessarily about the same topic -- I meant it very open ended. It's just that you came into the fray with such a specific piece that I wondered if you had any more like that. If it's too broad of a scope though, lol, pardon me :P Viva la horde, ~ GoatLordServant^(Talk) 17:45, 19 February 2026 (UTC)

No problem at all, sorry I can't be more helpful. Gnomingstuff (talk) 16:46, 20 February 2026 (UTC)

It's okay :) Viva la horde, ~ GoatLordServant^(Talk) 17:03, 20 February 2026 (UTC)

Could you add some examples? I'm sure you have them. sapphaline (talk) 20:12, 19 February 2026 (UTC)

If you still want more "wild" samples of text confirmed as generated by Claude: Diff/1339223651.

Claude is AI and can make mistakes.

The whole diff is an... interesting read. --Gurkubondinn (talk) 21:14, 19 February 2026 (UTC)

unironically the most transparent I've ever seen anyone be about their AI process here, accidental or not Gnomingstuff (talk) 21:33, 19 February 2026 (UTC)

Hi Gnomingstuff, I remember there was a discussion which went through a number of expand newcomer task examples and showed a lot being rapidfire done through llm expansion, generally to poor quality. However, I've not been able to find it, even looking through the archives of newcomer tasks and of AI cleanup. Possible I missed it, but do you remember this discussion? Looking to make a short proposal at Wikipedia talk:Growth Team features to disable that task. All it takes is an admin clicking the button under "Expand short articles" at Special:CommunityConfiguration. I know you have some concerns about other tasks, but I'd like to propose this one in isolation as it seems a particular locus of problems. CMD (talk) 16:10, 20 February 2026 (UTC)

Do you mean this? That's the only other one I'm aware of that I was involved in. I agree that the expansion task probably attracts the largest volume of text (although in practice people kind of blur them together).

I can also audit the more recent expand tasks (it's about 2 weeks' worth) -- I hadn't done it at the time because I ragequit after 60 when I realized I'd been copying the wrong diff number the whole time and had to go back and fix them all. Gnomingstuff (talk) 16:28, 20 February 2026 (UTC)

60 is probably enough, although not sure if it is needed if it requires more time. I raised this after coming across both the WMF discussion and another llm-expand I reverted. I'll start the discussion. CMD (talk) 07:20, 21 February 2026 (UTC)

Hi. This is in reference to the AI generated content on the page of Nagari Pracharini Sabha. This, perhaps, arise from my edits during Nov 2024, particularly in the 'Key Activities' section, but, (I also left a message on the page's talk page), none of the information is hallucinated, everything is sourced from a magazine (in Hindi), if you wish, I can share the pages of the mag over email. The 15th reference (as on 25th Feb, 2026) is about the same magazine. Other content on the page has been moved or expanded, and, I believe, that can't come under AI generated content. Vyom.Y (talk) 09:57, 25 February 2026 (UTC)

 You are invited to join the discussion at Talk:Illegal immigration § Requested move 25 February 2026. Edittttor (talk) 20:37, 25 February 2026 (UTC)

Why am I pinged here? Gnomingstuff (talk) 20:42, 25 February 2026 (UTC)

hey, we wrote the whole page again, which part suggests that it was LLM, let me know! MateenSalimBhat (talk) 05:24, 28 February 2026 (UTC)

It would help to leave the world of suggestions and just say, definitively, whether AI tools were used (including editing tools like Grammarly or translation tools like DeepL) Gnomingstuff (talk) 05:49, 28 February 2026 (UTC)

in his first draft yes which was months ago, in the next one, no! MateenSalimBhat (talk) 05:50, 28 February 2026 (UTC)

Hi Gnomingstuff. I saw your comment, "Newcomer tasks once again". I haven't been paying attention to these newcomer tasks. Am I assuming correctly that you are indicating that there have been similar problems with editors taking on newcomer tasks by using ai's inappropriately? Have there been discussions on content-related noticeboards about such problems that you're aware? I steer newcomers away from contentious topics, and think everyone, including Growth Team, should do the same. --Hipal (talk) 18:55, 28 February 2026 (UTC)

There's an open discussion on the Growth Team features talk page about it, plus a few on village pump, but the gist is that the feature does encourage -- in practice if not in intent -- people to make a lot of poorly reviewed or unreviewed edits at high speed (like 2-3 minutes between edits high). This has been the case for years but has accelerated since AI makes it easy to pump out edits at high speed.

I was probably unclear about my point though. My point is more "you can't really blame someone for doing something they were encouraged to do." This happens a lot (particularly with the expand/copyedit/update tasks), and it's not because there are hundreds of bad-faith malicious editors, it's because the feature provides a perverse incentive. Gnomingstuff (talk) 19:36, 28 February 2026 (UTC)

Thanks for the explanation. That makes sense. Steering editors in the right direction should be easy. --Hipal (talk) 19:53, 28 February 2026 (UTC)

Hello Gnomingstuff, I saw that you added the "AI-generated" template to the Airat Ichmouratov page. I really appreciate your experienced feedback in helping improve this article. I was wondering if you could recommend what exactly needs to be improved on the page. I would be happy to work on improving the article’s quality. I would greatly appreciate any advice or guidance you could offer, perhaps with some specific details about what could be improved. This template is new to me, and I didn’t encounter this issue in the past, so any suggestions would be very helpful. I have always hired a professional certified translator to proofread the texts before adding anything to Wikipedia, but it seems possible that AI may have been used in the proofreading process. In any case, as I mentioned, any help in pointing out the parts that could be improved would be greatly appreciated. Thank you in advance. Patrick0506 (talk) 00:28, 9 March 2026 (UTC)

Hi - the first thing I would do is try to get some clarity from the person proofreading it about their process. Gnomingstuff (talk) 20:43, 9 March 2026 (UTC)

Yes, I already asked about this and received confirmation that i was screwed, the proofreader mentioned AI had been used "very lightly" during the proofreading process. However, since this was done three years ago, unfortunately neither of us kept the original texts. Could you recommend what I could do to improve the article? I would be very grateful for any guidance. Thank you again for quick answer. Patrick0506 (talk) 22:27, 9 March 2026 (UTC)

Hey! You suggested I used AI for my edits of the Checkstyle article, and chose to delete the whole section. I spent a fair bit of time researching my changes and carefully reviewing the generated content before posting it, striving to resolve both the `stub` issue on the article and the `primary sources` issue.

After briefly reviewing your work, I noticed that for most articles you appear to tag them with the AI template, but in this case you reverted the content outright with the edit summary “rv to pre-AI version.” I found that surprising, because if the issue were with the sources I would expect a note about sourcing, if neutrality were the concern I would expect a note about neutrality, and if the prose were the concern I would expect edits adjusting the wording. Removing the content entirely seemed like the least productive approach in this situation.

Perhaps you noticed that this is my only contribution and assumed the content had not been reviewed. I would question whether that is a good assumption. I could understand taking a stronger action if the edits came from an anonymous account or a pattern of unreviewed AI contributions.

More broadly, I have noticed that several editors on your talk page have raised concerns about false positives when identifying AI-written text. From your earlier comments here, it seems your detection process often relies on stylistic, vocabulary, or tone-based indicators. In some cases the standard appears to be closer to “it feels like AI” even when there is no specific issue identified.

My concern is that style-based suspicion alone may not be enough, especially now that many editors may spend significant effort researching sources and verifying facts while still using AI tools to help with phrasing or structure. In those cases, the more important question seems to be whether the text introduces hallucinations, misrepresents sources, or otherwise violates content policies.

In the case of the Checkstyle article specifically, I wonder whether the tradeoff is worthwhile. Even if some of the reverted content may have involved AI-assisted drafting, reverting the article also: reintroduced the stub issue, worsened the primary sources issue, and removed a significant number of references.

Would it make sense in cases like this to take a lighter-touch approach first—for example, tagging the article for possible AI-generated content and outlining the specific concerns so that other editors can review it as well? That seems like it would preserve potentially useful sourced material while still addressing the risk of AI-related issues.

I want to emphasize that I am not opposed to efforts to clean up unreviewed AI content. I simply think the focus should be on verifiability and source integrity rather than prose style alone, and that reverting large amounts of sourced content should probably be a last resort rather than a first step. Jpstrommen95 (talk) 20:03, 10 March 2026 (UTC)

Hey! You suggested I used AI for my edits of the Checkstyle article, and chose to delete the whole section. I spent a fair bit of time researching my changes and carefully reviewing the generated content

...so you're complaining that I said the edits were generated by AI when they were, in fact, generated by AI? Gnomingstuff (talk) 23:12, 10 March 2026 (UTC)

I’m not disputing that AI tools were involved in drafting the text. My concern is less about the label and more about the underlying assumption that AI involvement alone justifies removing the content.

There is currently a strong stigma in many discussions that equates “AI-assisted” with plagiarism, fabrication, or unreliable content. However, the guidance at https://en.wikipedia.org/wiki/Wikipedia%3AAI-generated_content can essentially be summarized as: computer tools may assist in generating text, but editors remain fully responsible for verifying the accuracy, sourcing, and policy compliance of anything they add.

With that in mind, when I made the edits to the Checkstyle article, I made a deliberate effort to ensure that the claims were supported by sources and that the content aligned with core policies such as verifiability and neutrality. I fully recognize that my contribution may still have shortcomings, and I welcome other editors improving or refining the text. That kind of collaborative editing is the normal process here.

What concerns me is the idea of dismissing or reverting large amounts of sourced content primarily because of the method used to generate the prose. Ultimately, the key question should be whether the claims are supported by reliable sources and accurately represented in the article.

I’m also curious about where you personally draw the line when evaluating AI-assisted edits. In particular:

• What principles do you use to determine when AI assistance is acceptable versus when it warrants removal of content?
• Do you focus primarily on stylistic signals, or on source-to-text integrity and verifiability issues?
• In cases where the sourcing appears sound, would it generally be preferable to revise the prose rather than revert the content entirely?

I’m asking these questions in good faith, and—seeing that you are a very active and experienced editor—I value your perspective on these emerging issues that many of us are now navigating as AI-assisted tools become more common in editing workflows. Jpstrommen95 (talk) 23:43, 10 March 2026 (UTC)

The claims are not supported by WP:reliable sources. Most are interpretations of github code (WP:UGC even when not actual code), and one paragraph doesn't have a source at all. CMD (talk) 03:21, 11 March 2026 (UTC)

Thank you for the clarification, @Chipmunkdavis. I’ve reviewed the feedback regarding WP:reliable sources and WP:UGC. To keep this discussion focused on user behavior rather than a single article’s technicalities, I will accept the removal of the Checkstyle content for now and move that specific sourcing discussion to the Talk:Checkstyle page, where it can be addressed by the broader community.

In closing on this page, however, I want to address the broader pattern of mass deletions occurring here.

While I acknowledge the technical concerns raised in my specific case, there is a visible trend on this talk page across multiple users and domains. It appears that @Gnomingstuff is frequently reverting content based solely on the suspicion or label of AI usage, rather than identifying specific policy violations.

I would like to invite @Gnomingstuff to reconsider this "mass deletion" approach. While AI-assisted content requires scrutiny, it should be judged by the same standards as any other contribution:

- Verifiability: Is the information accurate?

- Neutrality: Is the tone appropriate?

- Sourcing: Are the citations reliable?

Assuming a policy violation exists simply because a tool was used—without verifying the actual content—risks removing valid, helpful information and can be seen as disruptive. I hope that future reversions will be based on specific, demonstrable policy failures rather than the presumed origin of the text. Jpstrommen95 (talk) 04:03, 11 March 2026 (UTC)

Are you using AI for your comments here? Gnomingstuff (talk) 04:29, 11 March 2026 (UTC)

Hi, sorry to bother you. I've restored your AI-generated tag on the article and rv'd the latest edit after an editor that I think we can all agree is using an LLM (see their AfC contributions too) has made significant changes to the article. However when you first added the tag you said "since v1" and having examined your conversation with the creator of the page I'm unsure if you're still of the opinion that v1 is also AI generated? Trying to figure out what version to best revert to. Fermiboson (talk) 20:31, 12 March 2026 (UTC)

Hi, thanks for the note -- by v1 I mean the first version of the article, from February 2025. Not sure what conversation you're referring to but I think this is fairly clearly AI-generated. Gnomingstuff (talk) 20:33, 12 March 2026 (UTC)

I was referring to User talk:Particleshow22#AI. Thanks for clarifying - I guess I'll just leave as is for now given there's no good version to revert to. Fermiboson (talk) 20:42, 12 March 2026 (UTC)

“AI version”? What do you mean by this???? On the edit of Kenneth Bøgh Andersen? I didn’t use any AI. I HATE AI. Luka1184 (talk) 11:16, 13 March 2026 (UTC)

The edit not only shows signs of AI use, but the very first reference I checked contained a hallucination: [Michael Grant...] contacted Andersen, praising it as "Thrilling, fascinating, profound and still somehow funny all the way through," highlighting its appeal in blending dark fantasy with lighthearted yet somewhat dark humor elements. The "highlighting its appeal" bit is AI slop that is not even remotely mentioned in the source, which just shares that Michael Grant said that comment and does not mention dark fantasy, blending fantasy with humor, whether the humor is lighthearted or dark, or really "highlighting" anything at all. There is no possible way that someone who read the source -- which is a three-sentence Facebook post -- can get that interpretation from that source. However, it is the same editorializing that AI does about virtually every source. Gnomingstuff (talk) 12:59, 13 March 2026 (UTC)

Hi, please refer to Talk:Traumatic Livelihood page regarding yr tagging of article as AI-generated, which is false (i have provided links to my sandbox page); would still like comment on how to write better though since yr writing critiques are valid. // Chchcheckit (talk) 13:34, 13 March 2026 (UTC)

A curious statement regarding the MACS MTO page : This article may incorporate text from a large language model. It may include hallucinated information, copyright violations, claims not verified in cited sources, original research, or fictitious references. Any such material should be removed. The reason given is: since v1 in 2025; note WP:AISIGNS particularly promotional tone, vocab distribution typical of 2025 LLMs etc.; also notice weird ref plaintext formatting in first diff (February 2026)

Have you read the sources cited in the References? HenriDavel (talk) 21:17, 17 March 2026 (UTC)

Did you use AI for this? Yes or no. Gnomingstuff (talk) 21:31, 17 March 2026 (UTC)

Misclicked rollback EvergreenFir (talk) 21:51, 19 March 2026 (UTC)

not sure what this is in reference to but no need to apologize! it happens Gnomingstuff (talk) 21:52, 19 March 2026 (UTC)

(talk page stalker) just a heads up, there’s a tickbox in your preferences that forces you to confirm a rollback with a second click instead of it being a direct link :) Danners430 _{tweaks made} 21:53, 19 March 2026 (UTC)

@Danners430 I might need that, thanks. EvergreenFir (talk) 22:01, 19 March 2026 (UTC)

I like your style Gnomingstuff, keep getting rid of AI stuff :) Hypawthetically (talk) 01:28, 20 March 2026 (UTC)