User talk:Pythoncoder/Archive 8

Hi Python,

Thanks for your message about whether the page was LLM generated. It wasn't - I just tried to follow general Wikipedia formatting. In fact, I tried to use AI to format the references and it couldn't do it. Are you able to further review the page I've made? Would really appreciate it!

Thanks Clapham12345 (talk) 06:16, 1 November 2025 (UTC)

Thanks for the clarification. I was on the fence so I appreciate your confirmation that the draft was human-written, and I'll remove my earlier comment from the draft. —pythoncoder (talk | contribs) 10:27, 1 November 2025 (UTC)

Hello, I was able to remove the citation errors on my article and resubmit it. You may find them here: Draft:Battle of Paris (Tennessee) 108.59.60.242 (talk) 14:14, 3 November 2025 (UTC)

Dear Pythoncoder,

The article has been enhanced based on your constructive suggestions. Appreciate if you could help to review the draft. Thank you. Profwinedit (talk) 18:32, 3 November 2025 (UTC)

Hello @Pythoncoder{{Ping|Pythoncoder}}

I would like to know what is it about the draft that reads like an advertisement? The page I wrote focused on the corporate information, similar how many other companies have done so far. We used known and trusted sources such as PR Newswire, USA Today, and similar. I am looking forward to hearing your input.

Trpk93 (talk) Trpk93 (talk) 13:16, 6 November 2025 (UTC)

PR Newswire is not a reliable source. It’s a site for publishing press releases, which are neither neutral nor independent. In fact, most of the sources you use are press releases, which do not count towards notability. In terms of the draft’s prose, if you can’t see why it’s nothing more than an advertisement for the company, then you shouldn’t be writing this article, because neutrality is a requirement for Wikipedia articles. —pythoncoder (talk | contribs) 13:42, 6 November 2025 (UTC)

Also, who is the “we” in the above sentence “We used trusted sources…”? —pythoncoder (talk | contribs) 13:44, 6 November 2025 (UTC)

My employer and I is who the "we" is referring to, since I don't have all the necessary information about the company, the employer was helping me put the article together. I've tried making it sound less like an ad, and I've replaced the biased sources with independent ones. Trpk93 (talk) 09:35, 13 November 2025 (UTC)

Got it —pythoncoder (talk | contribs) 10:03, 13 November 2025 (UTC)

Hi, a user added a userbox then, another deleted it and says "you need rollback rights" and idk why i need that but, they told me to tell you to help me. Tester6462656543 04:06, 7 November 2025 (UTC)

Context, they added {{User AntiVandal}}, but they did not, in fact, use AntiVandal. The other user did not tell tester to ask you, but they added the statement themselves, violating WP:TPO. --pro-anti-air ––>(talk)<–– 04:29, 7 November 2025 (UTC)

Aaaaaand they’re indeffed —pythoncoder (talk | contribs) 07:53, 7 November 2025 (UTC)

Hello @Pythoncoder:, I have tried to rejig the patchwork roads article based on your feedback. If it currently does not meet the guidelines, could you expand on the specific reasons why not? I'm using an LLM to convert my writing from Markdown to the Wiki Format to save some time (background in computer science!). Is that the issue? I was keen not to have to learn another wiki format, if possible, as I already get them all muddled up in my head, so I try to stick with markdown and then post-process them. I have also streamlined it to the most relevant part (media engagement) to save time, since I spent a considerable amount of time on the initial version.

Link: https://en.wikipedia.org/wiki/Draft:Patchwork_Roads

Liamjberrisford (talk) 08:16, 7 November 2025 (UTC)

Hello @Pythoncoder:, Would you mind expand on the rejection of the Patchwork Roads article for LLM use? Im a little lost of what changes are required to bring it into standard. Liamjberrisford (talk) 08:37, 7 November 2025 (UTC)

Hello @Pythoncoder:, I think the article might have been rejected for the incorrect reason. You initially rejected for LLM use, that I spent some time fixing and making quite substantial changes. Then someone else @Aesurias: rejected it for LLM use again, and so i resubmitted as it wasnt clear what changes were needed and then you changed your second rejection from LLM use, to a resubmit without changes. I think with both you and Aesurias reviewing wires are getting crossed. Would it be possible for only one of you to review so the goal posts dont keep moving? Im happy to make the needed changes, im just unsure at this moment in time what they are. Im not sure how it is LLM use when i wrote the content, went out and took the picture as an example etc. Im finding this process slightly frustrating but would really like to contribute. Liamjberrisford (talk) 08:43, 7 November 2025 (UTC)

Would it be possible to not use AI in your drafts so they are not declined for AI usage?

The ''usage in media'' section is clearly written by a large language model. Unnecessary use of bold text is another indicator.

The 'goal posts' have not moved, and there was no incorrect reason for rejection. Your draft was declined (for the third time) due to the fact that you resubmitted it almost immediately without making any changes. If you really want to have the article approved, I would scrap it entirely and start again in your own writing. Aesurias (talk) 08:48, 7 November 2025 (UTC)

Hello @Aesurias:, just FYI i think your first sentence is quite rude, and unneeded at this point in time. Im trying to engage in good faith. RE the 'usage in media', could you expand on what particularly is rather than just "vibes"? Thank you for the suggestion on the bolding. It's something that i have found helpful, but I am more than happy to remove it. If you have any other actionable steps that would help move it from "AI" generated then I would really appreciate them as i can fix them.

Is bolding the only thing that needs to be changed before resubmission?

Liamjberrisford (talk) 08:53, 7 November 2025 (UTC)

@Aesurias: I would also be happy to completely scrap it and start again if i was confident that it would then be fairly assessed. However given the current approach you have taken it feels that only very etheral reasons are given that are very hand wavy that you believe AI has been used. I think having very actionable changes is the best way forward so we can move towards something acceptable if that is agreeable to everyone. Liamjberrisford (talk) 08:59, 7 November 2025 (UTC)

Based on your previous messages on this talk page, I will assume that the reason it reads like AI is because you're converting it from Markdown. I think it would be advisable for you to just learn how to use the Wiki Format because the article reads as if it was created, almost entirely (with the except of the lede paragraph) by AI.

The entire 'usage in media' section reads like AI.

A human would specify which councillor and in which newspaper, but AI would write "A councillor quoted in a regional news article".

"A June 2024 article" is also highly suspicious and reads like AI-generated text. Aesurias (talk) 09:00, 7 November 2025 (UTC)

Also see:

"Several North American studies..."

"Some authorities..."

"Documented contributors include..."

"A UK resurfacing firm used the term in a blog post..." Aesurias (talk) 09:01, 7 November 2025 (UTC)

Hello @Aesurias: (also CC @Pythoncoder:), thank you for those suggestions; they help immensely. I can totally see what you mean now. I initially thought the opening paragraph was the issue. I think the problem was highlighted particularly in the usage in the media section, as with this being my first Wikipedia article, I think I strayed too far into being "neutral" and I think it resulted in me coming across as robotic. The usage in media is also where my interest in the topic stems from and so also plays a large part in the overcorrection. It is also slightly concerning to me that my formal sentence composition feels quite artificial, and is definitely something for me to keep in mind in the future.

I have now implemented all of the requested changes. I have also submitted it for review again (maybe a bit preemptively!), but if there are any further changes I can make, please do let me know and I will action them asap. Liamjberrisford (talk) 09:44, 7 November 2025 (UTC)

Hi @Pythoncoder

Thank you for taking the time to review the Draft:Class Saathi article. I understand the concern about the content appearing to be generated by an LLM. However, I’d like to clarify that the current version of the article has been completely rewritten and refined by our own content team members, without the use of any AI tools.

We have focused on ensuring that:

1. The content is entirely original and written in a neutral, encyclopedic tone.
2. All statements are backed by reliable and verifiable sources.
3. Promotional or non-neutral language has been carefully removed.

We’d be grateful if you could please take another look or advise if there are specific sections that still seem problematic. We’re happy to make any further edits needed to align with Wikipedia’s standards.

Thank you for your time and guidance. Tag-hive-new (talk) 10:32, 7 November 2025 (UTC)

I noticed that you once commented on an AfC submission that Whoever asked a LLM to write an article about this person should probably read up on the environmental impacts of generative AI, which does not seem civil. Additionally, the creator claims that their writing might look AI-generated at times, although I will grant that you probably did not know that as the claim appears to have been made after your comment. I believe you are acting in good faith, but without civility. Nononsense101 (talk) 21:35, 7 November 2025 (UTC)

Noted. I have reverted the decline+comment. —pythoncoder (talk | contribs) 21:38, 7 November 2025 (UTC)

Hello @Pythoncoder You have declined the page Swami Narasimhananda on the grounds that it sounds like AI generated. I had worked on this article for many days and all the sources and citations have been thoroughly verified. I need your help in pointing out the exact places where there are issues. I look forward to your experienced insights so that I can fix the issues. Please help me. Bindusouhrudam (talk) 08:59, 9 November 2025 (UTC)

Did you use AI or not? —pythoncoder (talk | contribs) 12:41, 9 November 2025 (UTC)

No. I have not used AI. I have searched the internet for reliable sources of data. However, I have changed the language now, and brought three independent sources to the beginning of the article. I have also deleted many citations and some content without citation. Also, I have removed the two photos I had uploaded earlier and replaced them with a copyright free image. Bindusouhrudam (talk) 13:03, 9 November 2025 (UTC)

I do not believe your claim that you did not use AI to write the revision of the draft that I declined. It contained multiple mistakes in formatting and style that are made almost exclusively made by large language models. —pythoncoder (talk | contribs) 14:46, 9 November 2025 (UTC)

That was written by me. Could you please point out the places that appear like AI so that I could fix them? Also, only facts have been written that are available from various sources. Bindusouhrudam (talk) 18:02, 9 November 2025 (UTC)

I’m also having trouble coming up with an explanation for the discrepancy in grammatical quality between the draft (where every revision is extremely professionally written) and your talk page comments (which contain occasional grammar mistakes). Also, why do some portions of the draft use correct italic formatting while others feature uncorrected Markdown? —pythoncoder (talk | contribs) 18:37, 9 November 2025 (UTC)

That is because I have written and rewritten the draft using my desktop and carefully gone through it while the talk page remarks I have written mostly from my mobile phone. I have spent many days in collating and verifying the data. It is all based on what is available in reliable sources. Bindusouhrudam (talk) 02:05, 10 November 2025 (UTC)

I am teacher and actively work against the use of LLMs. I daily face problems with student assignments because of LLMs. I appreciate your stand on LLMs, but I haven't used LLMs in this page. This is my first article in Wikipedia and I need the help of established editors like you. It took me much time to write this article but I have shortened it in the revision. Bindusouhrudam (talk) 02:13, 10 November 2025 (UTC)

@PythoncoderI do not see any of the signs of LLM writing that you have claimed. Nononsense101 (talk) 02:34, 10 November 2025 (UTC)

@Nononsense101 why do you say that? Frankly this is a black and white situation. The original draft is written in perfectly grammatical English, with many signs of LLM-generated content, while the author's talk messages here show a distinctly different style, with occasional grammatical errors. NicheSports (talk) 03:58, 10 November 2025 (UTC)

That was because I did not read the bottom section. That section shows a lot more signs of AI writing, including the list formatting. Nononsense101 (talk) 18:04, 10 November 2025 (UTC)

Every thing that I have written is backed by reliable source and just clicking the links will show this. Bindusouhrudam (talk) 18:08, 9 November 2025 (UTC)

Hello @Pythoncoder , I’m the author of the draft article “M-red Function” which you recently reviewed and declined. Thank you very much for taking the time to review it. I’d like to know exactly what changes are needed to make the article meet Wikipedia’s standards and be eligible for acceptance.

In particular, I would appreciate it if you could clarify what types of sources are considered acceptable. For example, would research websites such as Zenodo or ResearchGate, or analytical articles on independent scientific websites, qualify as reliable independent sources?

Thank you in advance for your guidance.

Best regards, M-red-real M-red-real (talk) 10:30, 9 November 2025 (UTC)

I could go reference-by-reference and explain why none of the current sources count for notability, but first I want a yes-or-no answer to the following question: Did you use a large language model in any capacity while writing Draft:M-red Function? —pythoncoder (talk | contribs) 12:49, 9 November 2025 (UTC)

Hello @Pythoncoder

Yes, I used the ChatGPT AI model to help draft and translate the text (because English is not my native language), but all the information in the article was collected by me personally (unless a translation mistake by the AI caused any errors).

If needed, I can completely rewrite the article and ensure all translations are accurate (although I currently believe the AI did not introduce any mistakes).

Also, could you please advise me on what to do regarding sources?

Thank you for your guidance,

M-red M-red-real (talk) 13:04, 9 November 2025 (UTC)

1. The use of unedited machine translations is not allowed on the English Wikipedia (see WP:MACHINE).
2. The use of LLMs to write articles here is very likely to be banned within the next month or so.
3. Just to cover all the bases, the use of LLMs to write talk page comments is already not allowed.
4. It doesn't matter whether you collected the underlying information yourself; it is still possible for LLMs to fabricate information even if this is true.
5. To be clear, if you wish to "completely rewrite the article", that should mean "rewriting it 100% by yourself without LLM 'help'", as opposed to telling ChatGPT, "Please rewrite this article and ensure all translations are accurate."
6. Regarding your sources: #1 is an article from another wiki that you appear to have generated and submitted yourself; it cannot be used as a source. #2 appears to be self-published as well and is thus unusable; it doesn't matter that it appeared on Zenodo as that website does not do peer-reviewing or fact-checking. #3 and #4 are personal websites and don't count for notability either.
7. This means that even if you did rewrite the draft from the ground up, it probably wouldn't be accepted (especially given your conflict of interest), because as it stands right now there do not appear to be any sources that are usable to prove the subject is notable.

—pythoncoder (talk | contribs) 14:40, 9 November 2025 (UTC)

Hello @Pythoncoder

I would like some guidance on how I can obtain independent sources. The challenge I face is a bit circular: independent sources need to recognize the function, but recognition of the function itself requires independent sources. I would really appreciate any advice you can provide on this.

Also, English is not my first language, but I am able to read and understand it, and I carefully check all content translated by AI to ensure the meaning is accurate.

Thank you very much for your help!Hello @Pythoncoder M-red-real (talk) 12:32, 16 November 2025 (UTC)

If you can’t find three independent sources that meet WP:42, then the subject is not notable enough for a Wikipedia article. Sorry. —pythoncoder (talk | contribs) 15:03, 16 November 2025 (UTC)

Hello @Pythoncoder,

Thank you for your guidance. I will do my best to find independent sources. M-red-real (talk) 09:45, 17 November 2025 (UTC)

Hello colleague, you rejected my article. I submitted it again for creation. I added three sources in English for your convenience. Could you please take a look at them

Link to my article:

https://en.wikipedia.org/wiki/Draft:PMC_Redan Bogdan Smailov (talk) 16:37, 9 November 2025 (UTC)

Colleague, I'm sorry, it looks like I added two or one. In Russian Bogdan Smailov (talk) 16:44, 9 November 2025 (UTC)

Hello and thank you for your review and help monitoring new pages! I left a comment in the talk page for that draft, but I am not THAT Juliet Shen. I have the same name but I work in a totally different field (online trust and safety) and discovered the other Juliet Shen's work from sharing the same name and years of "competing" for a digital footprint. We are not the same person, and I believe her work in fine art and type design is far more notable than anything I have done. Shensationalshen (talk) 20:41, 10 November 2025 (UTC)

Sorry about that, I didn't see that message. —pythoncoder (talk | contribs) 22:18, 10 November 2025 (UTC)

no worries, thank you again for the review and message ❤️ Shensationalshen (talk) 22:30, 10 November 2025 (UTC)

Good morning.

I see you rejected the article, but to be fair and with all due respect as I am editing for the first time, your rejection is more vague than anything in the content.

Would you be able to point out in the article examples of what may make it vague or incorrect, or essay like? regardless of how the article was created or written, I don't see in the article where the guidelines have been violated. Sources are correct, the information is not promotional, and everything is factual.

If there was a bit of actual guidance on what in the article needs to be fixed or what in the content violates the guidlines that would be a lot more helpful.

The reason for the rejection is very vague and doesn't point out exactly where the violations are in the content itself.

Any direct clues would be helpful please. ~2025-32914-37 (talk) 05:09, 12 November 2025 (UTC)

The draft is AI-generated and its writing style reads as such. It contains multiple issues with formatting and style. I suggest rewriting it from the ground up without using AI. The draft is short so it shouldn't take too long, and I guarantee you'll get a better result. —pythoncoder (talk | contribs) 08:01, 12 November 2025 (UTC)

Hi, Pythoncoder! I played the random draft roulette today and landed on Draft:Avea Life, which you previously declined for being AI generated. I think the most recent revision seems better, and the sources seem fine too. I'm ready to accept it, but since I'm still on probation, I'd like a second opinion if possible. Thanks! guninvalid (talk) 08:39, 12 November 2025 (UTC)

Nah, that draft still looks AI-generated. Note how many of the sentences are framed in terms of attributing the info to a media source even when a human would recognize such attribution as unnecessary. —pythoncoder (talk | contribs) 09:00, 12 November 2025 (UTC)

I do appreciate you checking in with me about the draft though. It’s always good to see more draft reviewers learning the ropes and helping reduce the endless backlog. Let me know if you have any other questions. —pythoncoder (talk | contribs) 09:02, 12 November 2025 (UTC)

I appreciate your time for reviewing my article. However, all of the information has been researched thoroughly including listening to the podcasts this subject has been featured in, these are also cited. Everything has been fact checked and the subject is a large well known figure in the beauty industry.

If you could provide any further advice it would be appreciated. ~2025-32991-80 (talk) 16:05, 12 November 2025 (UTC)

Thank you for adding additional sources to the draft. However, many of the sources are interviews, which generally don't count towards notability unless they are of the highest quality (as on e.g. 60 Minutes) — see Wikipedia:Interviews#Notability. EIN Presswire is a press release website; those don't count towards notability either. "The NYC Journal" looks like an unreliable source; I opened a few more random articles on that site and they were all credited to the same person, with no publication dates. Everything else is a primary source. If the subject is indeed well-known in the beauty industry, there should be some reliable, independent, secondary sources available about her, but right now the draft does not contain any.

This isn't necessarily the end of the road — Exodus Wallet was once a draft that I thought stood no chance of getting accepted, but I helped guide the author through finding sources to prove the subject was notable. So it's possible that there could be usable sources out there, but as it stands, even after your revision of the draft, the subject does not yet appear to be notable. —pythoncoder (talk | contribs) 17:16, 12 November 2025 (UTC)

Hi, Pythoncoder! Thanks for checking me earlier on the previous draft. I did want to ask if you could provide a second opinion on Draft:BGN Group. I almost want to accept it, and I think it meets WP:GNG based on the sources provided. A few sources are primary and several sources are pretty WP:RUNOFTHEMILL coverage, but I feel like there's enough to consider it notable, particular the NYT and Financial Times sources. Do you think it's is acceptable? guninvalid (talk) 20:16, 13 November 2025 (UTC)

That draft is definitely not acceptable. The draft is textbook LLM output — note the incorrect markup (Markdown instead of wikicode) in multiple sections. As such, I have declined the draft. —pythoncoder (talk | contribs) 20:47, 13 November 2025 (UTC)

Darn. Doesn't WP have an article on detecting AI output? Which WP article is that? guninvalid (talk) 22:13, 13 November 2025 (UTC)

Found it at WP:AISIGNS. guninvalid (talk) 22:52, 13 November 2025 (UTC)

Hi @Pythoncoder,

I hope you are doing well.

I’m contacting you regarding the Draft:Netis Groupe you previously rejected.

Since then, I have completely revised the article to address the issues: I improved the sourcing, ensured a neutral tone, and removed anything that could appear promotional.

However, because the draft was rejected (not just declined), the page is now locked for resubmission. Even after making significant changes, I am unable to submit it again.

Since I now understand that rejected drafts cannot be resubmitted, could you please advise me on the correct way to proceed if I want you to review the revised version?

Is there a specific process to request a re-evaluation or to show you the updated content?

Thank you very much for your time and guidance — I’m still learning the Wikipedia process and I want to make sure I follow the rules correctly. Clever-student25 (talk) 12:46, 14 November 2025 (UTC)

Congratulations on finally writing a talk page message yourself instead of copypasting from what a chatbot told you to say, as you were still doing as recently as 4 hours ago. Did you use AI in any capacity in the revisions you’ve submitted since I rejected your draft? —pythoncoder (talk | contribs) 13:21, 14 November 2025 (UTC)

Hi @Pythoncoder

Thanks so much for reaching out with your question. I really appreciate you taking the time.

To be completely upfront and honest with you: The most recent version of the draft was indeed written entirely by me, manually, after the initial rejection. I truly took your feedback to heart and made sure to craft the content, structure, and summaries myself.

However, I do want to clarify my use of AI during the earlier stages, as I understand how important transparency is for Wikipedia's policies. I used an AI tool in two specific ways: To help me brainstorm and discover potential external sources, which I then thoroughly checked myself. To assist with grammar and phrasing in English, as it's not my native language. search more about the sources to determine their reliability and neutrality

I now have a much clearer understanding of Wikipedia's AI policy, and I'm fully committed to ensuring that all future content I submit is my own original writing.

I'm very keen to get this article right, so please let me know if there's anything else that still needs adjustment. I'm ready and willing to make any further manual improvements required.

Thanks again for your patience and guidance.

Best regards, Clever-student25 (talk) 14:14, 14 November 2025 (UTC)

• Gen.J.S. Flunkinowski (talk · contribs)
  • Draft:List of Pamunkey rulers (edit | talk | history | links | watch | logs)

Gen.J.S. Flunkinowski (talk) 18:04, 14 November 2025 (UTC)

What is your request? —pythoncoder (talk | contribs) 19:44, 14 November 2025 (UTC)

I was very surprised reading through Draft:Program of the National Council of the Resistance. It's very well-written, especially compared to the slop I'm used to finding on random submission. Does this draft look acceptable? The prose looks good and the sources look... acceptable, even if I can't read them, and the copyvio turned up a 13.8% very unlikely. I did pick up on the use of "programme" and "nationalization" which are from different accents, but it seems the document itself uses both spellings so it's probably fine. guninvalid (talk) 09:42, 15 November 2025 (UTC)

The draft appears to be a translation from French Wikipedia — their page is at fr:Programme du Conseil national de la Résistance. These show up from time to time in AfC and they're generally pretty good. If you're satisfied with it, I think you can go ahead and accept it.

(FWIW "nationalization" is an acceptable variant spelling in British English — see Oxford spelling.) —pythoncoder (talk | contribs) 10:13, 15 November 2025 (UTC)

Draft:Disqualification of convicted representatives in France also seems to fall in this bin. I'll give them both another once-over tomorrow and probably accept them. Thanks for letting me pester you lol guninvalid (talk) 10:21, 15 November 2025 (UTC)

Hello, Thank you for your review. I have added two more primary sources as requested. Please let me know what else I can provide for this article to be approved. Warner4992 (talk) 21:47, 15 November 2025 (UTC)

Thank you for your time and review. I Have re edited the draft Draft:Connester according to the suggestions you made. Appreciate if you could check and if it is par with Wikipedia guidelines. Stingkygow (talk) 00:12, 19 November 2025 (UTC)

HI

Could you please approved this one

Draft:Bargaon Village , Saharsa Chandratimessaharsa (talk) 20:30, 24 November 2025 (UTC)

What have you done to address the issue that prompted me to decline your draft? —pythoncoder (talk | contribs) 20:40, 24 November 2025 (UTC)

i made the changes now , please approved Chandratimessaharsa (talk) 18:29, 25 November 2025 (UTC)

Hey, can you review Draft:Alaid Volcano? I need to get this to WAM ASAP. Thanks! Earth605^talk 17:03, 27 November 2025 (UTC)

I went ahead and accepted it. Thanks for improving Wikipedia! guninvalid (talk) 19:02, 27 November 2025 (UTC)

Hello @Pythoncoder,

I said this on the help section of the wiki but I would like to add that the page was created for my Grandad for is Scientific glassblower who has a lot of notable information about him, but unfortunately I made the mistake of using Ai to help write, which caused issues with the toning and sourcing. Would I be able resubmit if I rewrite and change it. I'd appreciate any advice on how to rewrite the draft (or even help with editing it) to fix any issues. Thank you! Wiki4RAMZY (talk) 17:49, 30 November 2025 (UTC)

First of all, I suggest that rather than attempting to fix the LLM output, you clear the existing content (except the previous comments from reviewers) and rewrite the draft from the ground up. Don’t include sections on “media coverage” or “context”. Be aware that since you have a conflict of interest, you may write nonneutral prose without realizing it, so don’t be afraid to keep asking questions like you’re doing right now. When you’re done rewriting the draft, please let me know so I can take a look at it. —pythoncoder (talk | contribs) 20:30, 30 November 2025 (UTC)

Thank you so much for the feedback @Pythoncoder i don't really know how i should start it now or change as im afraid of getting it wrong again. But I'll see what i can do. Thanks! Wiki4RAMZY (talk) 21:27, 30 November 2025 (UTC)

Hello @Pythoncoder,

Hope you're doing well, I'm about to rewrite the draft from scratch as you said. But before I start, I would just like to ask about the sourcing requirements when it comes to personal detail like age, date of birth and birthplace. Would I be able to add these as I'm the grandson or if its not stated by sources should I leave them out? Thank you for your guidance! Wiki4RAMZY (talk) 19:42, 2 December 2025 (UTC)

You will need a source for the age/DoB/birthplace info — see WP:NOR. Sorry. —pythoncoder (talk | contribs) 19:58, 2 December 2025 (UTC)

I found a podcast he did on BBC Radio 4 Inside science where he stated he left school when he was 15 in 1968 here's the podcast Does this count as Source as he will be 72 now? Wiki4RAMZY (talk) 20:29, 2 December 2025 (UTC)

Yes, that should work as a source for his age —pythoncoder (talk | contribs) 20:51, 2 December 2025 (UTC)

Hello @Pythoncoder hope you're well,

Could you recheck over my Draft please I've cleaned it up and only included Verifiable sources. Thank you! Wiki4RAMZY (talk) 20:00, 16 December 2025 (UTC)

Yes. I had help on writing and organizing the article to Wikipedia standards using Claude. I also edited the article using Grammarly to make it more readable. Yes, I want to create a new category for "Women in organized crime" because it seems bananas that there isn't one already. Is that a crime? So, are the "See also" links a red flag for the use of LLMs? I wish my brain were an LLM, but sadly, it's not. Those are my creations. The ones I listed, besides "Women in organized crime", are "Prohibition in the United States" and "Rum-running", which, if you read the article (which you clearly did not), make total sense and are the most obvious choices. Three total categories got you in a twist?

Your comment links to Wikipedia:Writing articles with large language models, but I am not sure you've either thoroughly read it or understand it. First, to quote you: "Wikipedia guidelines prohibit the use of LLMs to write articles." (which I've seen you copy/paste on many articles) Now, to quote the Wikipedia advisory: "Large language models should not be used to generate new Wikipedia articles from scratch."

"From scratch" are the operative words.

Here's what Claude did not do (or could not do): 1. Discover the topic of a fascinating person (of note!) by first reading a profile of her in the Greenfield (MA) Recorder 2. Research for 4 full days, primary sources, some behind a paywall 3. Dig through the National Archives for Court Journals dating from 1925-29. 4. Researching Hilda Stone's family history on familysearch.org

I see you @Pythoncoder. After reading this, I clocked you. I know what your deal is: you want to be the AI police on Wikipedia (which is funny because you are probably using your algorithm to analyze articles for submission). You've got an agenda with no sense of proportion, watching out for the "AI slop", but never actually reading the articles you are "reviewing".

Funny, broadly, I agree with your mission.

So, you start your introduction to me (a first-time article writer) with an accusation. No questions. You didn't make an effort to understand what happened and how it happened. Accuse first, ask questions later.

Whatever. I see you pride yourself on being a boss of sorts on Wikipedia. Good for you. I've got a life and wanted to share some information that I learned. I did a lot of research. A lot more than some of the reed-thin articles that I see on Wikipedia.

I've got the knowledge about Hilda Stone (and all my research materials), which makes me very happy. Getting approval from the WikiPolice means nothing to me. Delete the article! I. Don't. Care. Billhector (talk) 01:31, 2 December 2025 (UTC)

While I don't appreciate the unfounded accusations that I don't read drafts and I use algorithms for AI detection (neither of which are true), I have retracted my previous comment on your draft based on the additional information you provided here. My standard practice when I'm on the fence about whether a draft is AI-generated is to leave a comment rather than just declining the draft, though when I'm wrong, I prefer when submitters set the record straight calmly and politely rather than... doing what you just did. —pythoncoder (talk | contribs) 03:26, 2 December 2025 (UTC)

That was me being calm and polite when someone makes "unfounded accusations" at me. Just to add more context: I've got a graduate degree and know my way around primary sources. I know how to evaluate a source for reliability and validity. I know how to cross reference assertions in newspaper articles, like I did by checking the Boston American against the actual court records.

What, does my article have 40+ sources? Do you know many "AI slop" articles that have that many sources?

Do whatever you want. This will be my last Wiki article that I submit. I knew the horrible reputation of some (not all) editors of having a gatekeeping fetish, but I tried to contribute something good anyhow. That's on me for not trusting my instinct and not bothering.

I would say that I want to talk to the manager, but I don't care. Have fun on your Wikipedia. Later. Billhector (talk) 04:44, 2 December 2025 (UTC)

"I... know my way around primary sources." That's great, but Wikipedia is a WP: TERTIARY source. That means we need to cite WP:SECONDARY sources instead. guninvalid (talk) 07:33, 2 December 2025 (UTC)

"Im [sic] a dumbass"

Checks out.

You didn't read my article, so you're not sure how I used the primary sources. Read the article before jumping to conclusions. I'm not going to explain everything here.

You link to the definitions of tertiary and secondary sources. Either you didn't read them or don't understand them. Choose one.

In fact, primary sources are allowed. To quote:

1. Primary sources that have been reputably published may be used in Wikipedia, but only with care, because it is easy to misuse them.
2. Any interpretation of primary source material requires a reliable secondary source for that interpretation. While a primary source is generally the best source for its own contents, even over a summary of the primary source elsewhere, do not put undue weight on its contents.
3. A primary source may be used on Wikipedia only to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge. For example, an article about a musician may cite discographies and track listings published by the record label, and an article about a novel may cite passages to describe the plot, but any interpretation needs a secondary source.

To make it straightforward for you, court records (primary sources) that verify newspaper articles (secondary sources) are allowed. Unless you are contesting the validity of the federal record from 1926? Is that what you are doing?

No, you aren't doing that. You want to jump into a conversation without any understanding of the situation and ... troll?

Why are people who have never read my article commenting on it?

Man, I can't wait until this thing gets published (or not), then I can delete my account. Billhector (talk) 20:49, 3 December 2025 (UTC)

@Billhector, I am going to respectfully request that you speak to fellow editors civilly and try not to cast aspersions. You are very new and may not be aware of assuming good faith, which is the fourth pillar Wikipedia is built upon.

Your draft is strong, however has several sourcing issues. FamilySearch is not reliable (per WP:FAMILYSEARCH), due to being primarily user-generated. Many of your sources are also paywalled, which means verifying those sources will take more time. Your draft will eventually be reviewed and hopefully accepted! Happy editing! 11WB (talk) 21:29, 3 December 2025 (UTC)

I did not know that about familysearch. Thanks. Given that, while I searched FamilySearch for information about Hilda Stone and her family, I verified everything with the actual census documents. I don't think I made any assertions without the documents in hand.

Yeah, reading through the familysearch note, I used the primary source documents like the census the same way that I did the Federal Court Journals: verifying newspaper accounts of the date and locations of events. Billhector (talk) 22:24, 3 December 2025 (UTC)

Again, @Billhector, great work, and that would be great for a published paper or a joirnalistic profile. But that's also the definition of WP:SYNTH, which is a kind of WP:ORIGINALRESEARCH which is not allowed on Wikipedia. It would be great for submission to a historical journal, but not for Wikipedia. guninvalid (talk) 00:04, 4 December 2025 (UTC)

@GuninvalidWho are you? You are not an editor. You don't know what original research is, or primary, or tertiary, so who are you? I know you're a troll. Billhector (talk) 00:35, 4 December 2025 (UTC)

oh, and I forgot: you don't know what original research is either. Billhector (talk) 00:36, 4 December 2025 (UTC)

@Billhector, I remind you again of WP:CIVIL. This is a policy. 11WB (talk) 01:07, 4 December 2025 (UTC)

• Miamijane (talk · contribs)
  • Draft:David Oscar Markus (edit | talk | history | links | watch | logs)

Hi. You rejected my article on suspicion of use of AI. I assure you that AI was not used for a single word in the article. I am a trained lawyer and write well. This is a revised version of my prior submission of a significantly lengthier entry that I was told read too much like a biography. I have followed all guidance and suggestions about how to make the entry acceptable for inclusion. I totally understand the concern about use of AI, but this is not AI. Please reconsider.

Jane Whitmor (talk) 16:06, 3 December 2025 (UTC) Jane Whitmor

Jane Whitmor (talk) 16:06, 3 December 2025 (UTC)

After reading the draft again, I have concluded that my decline was incorrect, and I have reverted it. —pythoncoder (talk | contribs) 18:43, 3 December 2025 (UTC)

Hello. My article was rejected for the use of AI. Based on the advice from user 331dot in the AfC, I would like to humbly ask you to reconsider my previous submission. I have taken this feedback seriously and (*almost*) completely rewrote every sentence to go after that natural tone. I sincerely apologize for my previous use of AI in preparing this article. The content has been rewritten to basically focus on:

1. A quick history
2. The legal dispute
3. Positions of the org reported on by newspapers
4. The 2025 controversy with the group chat leak & how it was reported on by several media outlets

I am hopeful that this new version addresses the root concern. Would you be willing to take another look?

123quincyjeeds (talk) 18:49, 4 December 2025 (UTC)

Thank you for rewriting the draft. I have re-enabled the option for you to resubmit it. —pythoncoder (talk | contribs) 20:55, 4 December 2025 (UTC)

Just to confirm, is the issue message still saying that it shows signs of being written by an LLM? Or do I need to resubmit it beforehand. I've redone everything and am seeking clarification. Thanks!! 123quincyjeeds (talk) 20:57, 4 December 2025 (UTC)

I'm going to do another round of copyediting, and then I'll resubmit the draft on your behalf for someone else to review. —pythoncoder (talk | contribs) 20:59, 4 December 2025 (UTC)

Thank you so much!! 123quincyjeeds (talk) 21:06, 4 December 2025 (UTC)

 Done Draft has been resubmitted. We're currently in the middle of a draft backlog reduction drive, so it should be reviewed relatively quickly (i.e. in a few weeks rather than a few months), though I can't provide an exact date because we're all volunteers here. —pythoncoder (talk | contribs) 21:08, 4 December 2025 (UTC)

Looks like the page just got accepted —pythoncoder (talk | contribs) 21:24, 5 December 2025 (UTC)

Woohoo!! Thank you so much!! 123quincyjeeds (talk) 22:29, 5 December 2025 (UTC)

Hi Pythoncoder. I just wanted to continue our conversation from Draft:Alex Randall Kittredge. I absolutely appreciate your reviewing and think that the vast majority of your AI declines are obviously well-warranted, but I've been finding some of your declines a little hard to understand. It looks like you're giving flat "ai" declines due to stylistic AI signs that are sometimes easy to fix or that could have been created by a human editor, and that in my view are not strong evidence on their own of any problematic AI usage.

For instance, I agree that overattribution like "attracted international coverage in arts and general media" and "Goodreads lists several short works by him" could be a sign of LLM usage, but it's also common for human editors to write like that, especially given how much we drill into them at AfC that they need to focus on sources and explicitly show that their subject has attracted media coverage. It also seems like you're declining a lot of drafts solely for having red-linked categories or redlinks in their see also sections, which in addition to being a sign of AI usage can also be the result of things like interwiki translations, or can be done by human editors (as with the discussion regarding Draft:Hilda L. Stone above).

In the cases that are not blatantly obvious (i.e. where the draft is not raw AI slop), would you be able to try to leave a quick comment in future pointing out why you have declined a draft as LLM-generated? I think that pointing out the problems you have noticed (e.g. fictitious references, tone issues, failed verification, AI formatting) would help to avoid a lot of confusion from submitters and from other reviewers. MCE89 (talk) 16:24, 5 December 2025 (UTC)

I've been trying to list decline reasons more often recently, but I will be more consistent about it in the future. —pythoncoder (talk | contribs) 17:02, 5 December 2025 (UTC)

(Sorry to jump in.)

It's less about what is being said, than what is being said and how. For example, humans could write the phrase "profiled in". In practice, though, they aren't doing that very often, even when they received the same AfC drilling.

To illustrate:

• The dataset I'm working with currently has that phrase at roughly 2800% more frequent in AI-generated Wikipedia text after August 7, 2025 than in human-written Wikipedia text.
• But it doesn't seem likely to be an artifact of the dataset containing lots of drafts; it's roughly 540% more common in that post-August 7 AI text than pre-August 7 AI text, despite both being heavy on drafts...
• ...and when you compare the pre-August 7 AI text to the human text, there isn't a meaningful difference in frequency for that phrase at all.

(The date is a rough cutoff, corresponding to GPT-5's release date. I suspect this tic didn't actually start with GPT-5, but there are lots of LLM tools in the mix and who knows what is being used, so all that can really be said is "it only really started showing up with newer models.") Gnomingstuff (talk) 00:32, 13 December 2025 (UTC)