c:file:ExAdminMop.jpg appears to be AI generated. I would highly recommend refraining from using AI content on the Signpost where practical. Some Wikipedians have strong opinions on AI, and when it is used, it can detract from the content of the article. Mitchsavl (talk) 07:43, 11 March 2026 (UTC)[reply]
@Pine: There's gotta be something else we can use for that, right? There is going to be like 60kb of dung on the talk page if that runs. jp×g🗯️10:01, 11 March 2026 (UTC)[reply]
Hi @Mitchsavl and JPxG: while making changes to the gallery, I replaced that image in order to comply with JPxG's request, but I prefer the previous one. On principle, I don't object to using generated content if it's an improvement over alternatives and there aren't legal problems with using the generated content. By the way, Mitchsavl, please ping me if you have some feedback regarding a Signpost piece that I'm working on. I might not have seen this comment until much later if JPxG hadn't pinged me. Thanks, ↠Pine(✉)04:51, 13 March 2026 (UTC)[reply]
This was submitted in last issue, but not published for reasons including the author's use of AI. They have since revised the submission and put their own human review and intention backing it.
If anyone has feedback on this then please speak up. I think the topic and subject matter is interesting, and know it will attract reader comments. I wish that the focus of the article could be on its content and not the author's disclosure of the use of AI in writing it. Bluerasberry (talk)20:08, 17 March 2026 (UTC)[reply]
There was a recent major Wikimedia community decision about acceptable AI, and both apply to this piece. This decision recommends that authors can use AI to copyedit their own writing, which has happened here, and to assist with translation, which is also relevant to this piece. I would like to find an acceptable way to review this piece and publish it. Bluerasberry (talk)15:42, 25 March 2026 (UTC)[reply]
Could make a nice mention in the next news and notes. I found the line "reads like it could have been contributed by a graduate history student" interesting because I started editing Wikipedia as a graduate student. When I have gone to snoop user pages I have been surprised by the high number of regular editors with post-secondary education. I would think the majority of GA have been at least edited by someone with post-secondary education. Though I may be biased because I do not edit sports or celebrities. Czarking0 (talk) 03:01, 18 March 2026 (UTC)[reply]
The New York Times published a piece today on activity by Epstein's PR/reputation management team. I added it to In the media. Though it doesn't credit Smallbones or The Signpost, it almost reads like they used our recent coverage as a starting point. Bri.public (talk) 15:47, 18 March 2026 (UTC)[reply]
They do reference Smallbones/The Signpost, albeit via embedded link: Team members created networks of fake Wikipedia editing accounts, sometimes known as sock puppets, to sneak changes past administrators, whose accounts they also tried to disrupt by hacking.Wrackingtalk!21:52, 18 March 2026 (UTC)[reply]
Thanks for this @Bri and Wracking:. I saw the NYT article yesterday and have greatly modified my original opinion of it since seeing Wracking's comment on it here! Bri's write-up in ITM seems fair enough, I guess, but might also be modified. I'll add (for this talk page only) that almost everybody loves seeing some credit for their work in a paper like the NYT. It just affirms that what they are doing is worthwhile - which is all Signpost writers get "paid" for our contributions here. While complaints about not being credited are understandable, they are generally not worthwhile. Not all journalists have the time or skills needed to understand the contributions of others (see, e.g. an ITM item about The Times (of London) a couple of issues ago).
I do expect to have a Disinfo report for this issue, but there are still some remaining unusual difficulties left to be ironed out. It should be IMHO (!) one of my best.
I added a link to a Jerusalem Post story and a bare description of the subject, the IRGC. The story is about the IRGC-controlled media used as citations tens of thousands of times in four language versions of Wikipedia including enwp. The Post to their credit described the methodology well: they put together some kind of database query to get quantitative data on how many times 21 specific domains were cited. Unfortunately they did not list all the domains, but maybe they would talk to one of our reporters.
To do a spot check, I created a similar database query here, but with only five domains that were specifically mentioned. If we get the fuller list of domains, I'd be happy to update the database query.
It looks like the Reliable sources noticeboard has had a few discussions of IRGC reliability, none of them positive. I created another query here for use of over 100 IRGC-owned Internet domains that were seized by the FBI in 2020 and are still used as citations. ☆ Bri (talk) 00:10, 19 March 2026 (UTC)[reply]
@Bluerasberry, Bri, and JPxG: At the top of Itm I found Bluerasberry's story very confusing and offpoint. Wikipedia was only mentioned once in the 2 sources, but in the writeup it was the very center of the story. In short I changed it - please do whatever you think is best with it.
Y@Smallbones: Your edits are entirely an improvement. Right - "in the news" is not the conventional place to put a piece like this, and I inserted my own views or wiki advocacy beyond what the sources state. The original articles are about Internet Archive. I do not think it is overstretching to have our own reporting focus on Wikipedia, but Smallbones, your rewriting is much more aligned with the normal way of doing things than my text was. Bluerasberry (talk)23:42, 30 March 2026 (UTC)[reply]
If anyone has feedback on my submission then either edit it directly or share here. The attraction here is the videos but I also wrote text to present the wikiconfernece experience. Bluerasberry (talk)17:29, 26 March 2026 (UTC)[reply]
Yo Bluerasberry. Thanks for taking this on. The writing seems very clean and readable. There's a lot of text content, and I was actually a bit surprised when I scrolled past the videos and found it. Do you think presenting it a little differently might help prepare the reader for what's there? It almost feels like there's enough for more than one issue, but I'd understand if you want to keep it all together. ☆ Bri (talk) 01:26, 27 March 2026 (UTC)[reply]
@Bri: Thanks for the review. I just split it, and now looking at it, there is a lot of content.
I am indifferent about having both of these in the next issue, versus publishing videos in this issue and the text article in the next. You suggested publishing across issues - would you please make the call of all now versus now some, later more? It is priority to get videos out now though.
Looks great, I think the Community view was a good choice of venue for the videos. It is probably the case that separating your personal statements from the reflections of the other participants is probably preferable, as well. ☆ Bri (talk) 04:39, 28 March 2026 (UTC)[reply]
I've uploaded my submission to the Disinfo report and marked in Ready for copy editing. It's a bit long at about 2,500 words. If you have feedback on it, please e-mail me. — Preceding unsigned comment added by Smallbones (talk • contribs)
Noted that the paid-editing series sidebar was a bit out of date, so I manually updated the tags for a few of the sincewise written articles at Wikipedia:Wikipedia_Signpost/Author/Smallbones with paidediting. This should fill it out a little. If this is not what's desired, feel free to untag them, or whatever -- it's just the tags haven't been maintained in quite some time so they were falling behind. jp×g🗯️05:53, 31 March 2026 (UTC)[reply]
@Bri: Sorry for the last-ditch report, but since now we know that Wikinews will shut down permanently from May 1, I can add a brief blurb to the column, if you'd like to... or else, we can save it for the next issue and provide some more in-depth coverage. Let me know which option you like the most! Oltrepier (talk) 15:38, 30 March 2026 (UTC)[reply]
Let's do both! Can you put a brief mention in the current issue? Also can you quick look at newsroom archive 45 to see if the discussion links provided there are still useful to include? Bri.public (talk) 15:56, 30 March 2026 (UTC)[reply]
As usual, we are preparing this regular survey on recent academic research about Wikipedia, doubling as the Wikimedia Research Newsletter (now in its sixteenth volume). Help is welcome to review or summarize the many interesting items listed here, as are suggestions of other new research papers that haven't been covered yet. Regards, HaeB (talk) 09:41, 29 March 2026 (UTC)[reply]
The likelihood of me being available tomorrow for anything other than brief interstices is quite low. If anybody wants to publish tomorrow, I am happy with that, otherwise I will be available only on Monday. @Bri:jp×g🗯️09:49, 29 March 2026 (UTC)[reply]
I'll be in and out today but can check on status in the evening. Does Monday (U.S.) publication work better for you, JPxG? ☆ Bri (talk) 14:03, 29 March 2026 (UTC)[reply]
To clarify, the publication deadline had actually already been set (perhaps by mistake?) to Monday, i.e. now.
Btw, I'm running late with RR, so feel free to publish without it, but I should have it publishable in a few hours otherwise. Regards, HaeB (talk) 20:35, 30 March 2026 (UTC)[reply]
"Opinion" seemed a little sniffly so I went to one of those goddamned stupid sniffer websites and it said every sentence had GPTsmell except for, like, two. In the history I saw that those two were edited by Smallbones.
Wikipedians sure do be loving to accuse stuff of being slop, and yell at me for giant paragraphs upon paragraphs for being an evil piece of shit who hates America if we run anything that has slop in it, so if this is the case I don't want to run this. jp×g🗯️09:48, 31 March 2026 (UTC)[reply]
I am also, again, not running the Op-Ed because it is LLM stuff. I was very sternly convinced of this the last issue, and I guess the same article is just back. I really do not want to run this. Does anybody remember what happened the last time we ran some thing where a random person slopped their op-ed? Yeah, I know, it's their opinion so it doesn't make sense to say it isn't verifiable. Well, here is what happened: nobody gave a fuck. They just complained about it, at great length, and went to great extremes in declaring that the Signpost was cooked and chopped and washed and failing and dying, and everyone associated with it was a loser, and blah blah blah.
Mostly what I do on Wikipedia is log in every couple weeks and get screamed at for shit. I do not want to get screamed at for this. I do not give enough of a fuck anymore to get screamed at for this. I do not want to run a LLM op-ed for the sake of letting someone defend themselves when the main thing they're accused of is using LLMs in a way that pissed everybody off. Like, is this not the most obvious pulling-a-pin-out-of-a-grenade-and-shoving-it-down-your-pants thing ever?
I don't want to be anywhere near that crap when it goes off. I didn't want to be near the last one either, but there was just too much shit going on for me to close-read every submission (that had no actual thing wrong with it) and larp a noir detective by running it through a slopsniffer to joust at windmills. But this time I don't give a fuck anymore so I am just going to run every op-ed through a sniffer and if it ticks up too high then tough luck.
If someone cannot be arsed to just sit down and write a whole submission on their own, then I definitely cannot be arsed to agree as editor-in-chief to be subjected to a dozen and a half people trying to get the Signpost MfD'd over it and calling it a garbage rag sorry sack of slop shit that isn't worth the paper it isn't printed on for the money nobody pays. There is not time enough in the world to be the designated contact for this great wailing and gnashing of teeth over publishing a thing that could just as easily have been sat down and written by a person. I am not going to publish it. Please do not move it back to /Next issue/ because unless it is completely rewritten I am going to keep spiking it over and over until the sun burns out. jp×g🗯️09:58, 31 March 2026 (UTC)[reply]
JPxG, I support your decision to decide what goes through your approval. Maybe we should start formalizing that. As in, a checklist of what op-eds are acceptable and what are not. For anybody who didn't look at the draft, it was clearly disclosed by the author as drafted with assistance from an LLM. But even so, I think the E-in-C has the final say on what they personally approve. ☆ Bri (talk) 16:02, 31 March 2026 (UTC)[reply]
@JPxG: Thanks for the clear feedback. I appreciate the hierarchy of having an editor-in-chief make decisions, and I also appreciate the explanation, even though I personally am already ready to take no for an answer without explanation for any submission. Like Bri, I support formalization of giving the last word to the EiC, even without explanation or discussion, because I know the circumstances without being told, I know that I am pushing boundaries with this submission, and I know how short on time all of us are. I think we have a good collaborative system here that works well.
Other thing to formalize - no AI submissions in The Signpost, so if that is the rule, then I like that. If we were to make an exception, then I could support publishing non-English submissions with AI-translation, whenever someone who communicates best in another language has something to say to English Wikipedia readers of The Signpost.
Thanks JPxG for editorial commitment and please let me know what I can do to keep this fun for you. I hope that you can think of this decision you just made as entirely enjoyable, productive, and having a useful impact on quality of journalism. Bluerasberry (talk)16:19, 31 March 2026 (UTC)[reply]
Thank you for the explanation, even if the tone was rougher than I would have hoped for a submission that represented genuine effort and full transparency about its process.
I won't be resubmitting. I'm not willing to do a full human-only rewrite for the same reason I wouldn't write on a typewriter — LLMs are part of how I work, as they increasingly are for many contributors. If Wikipedia and its publications move toward rejecting that entirely, I think they will lose contributors, and I include myself in that.
I do want to support @Bluerasberry's and Bri's suggestion to formalize this as policy. Contributors deserve to know the rules upfront rather than spending time on submissions that will be spiked. A clear policy is better for everyone, including editors who don't want to be put in JPxG's position of having to make these calls ad hoc under pressure.
Looking at the calendar, we could do April 13 and 27. Otherwise, just one April issue. Thoughts?
As for myself, I'd rather steer clear of April 20 as a publishing date (or that weekend) as I'm signed up for the WP:420 collaboration (as is Bluerasberry). So if we just do one April issue, April 27 is preferable. Bri.public (talk) 16:59, 31 March 2026 (UTC)[reply]
I posted news from today about Serbian Wikipedia editors banned
This is a development in a complicated multi-year story. These banned editors are highly active, with at least one each highly active in English Wikipedia, Wikidata, and Commons. I linked previous Signpost coverage here and have this framed to include a general explanation of what bans are and what they mean.
Politico's Eulogy for CIA Factbook does not mention Wikipedia, but it feels very encyclopedia-adjacent. Maybe it can be worked in somewhere? The closure of this and threatened blocking of sites to Internet Archive so close together seem to augur … something. ☆ Bri (talk) 13:55, 6 April 2026 (UTC)[reply]
We had an AI submission and declined to run it. The person whose piece was declined requested that we clarify Signpost policy to decline AI, and I think that is a good idea.
I have this piece framed as "from the editors". I invite anyone to co-sign on this, and also anyone to edit any or all of this text, including deleting or changing it.