Wikipedia talk:Wikipedia Signpost/Single/2023-02-20

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia


Comments[edit]

The following is an automatically-generated compilation of all talk pages for the Signpost issue dated 2023-02-20. For general Signpost discussion, see Wikipedia talk:Signpost.

Cobwebs: Editorial: The loss of the moral high ground (19,524 bytes · 💬)[edit]

Discuss this story

"...original illustrations of children engaged in sexual acts... Sanger referred to an early 20th century colored illustration of a young girl performing oral sex on a much older man. Its caption reads: “If Mom returns? She'll tell you that it's very rude to talk with your mouth full." The image is accessible via Wikipedia's article on "Pedophilia," at the bottom of which is an image with a link directing readers: "Wikimedia Commons has media related to Pedophilia.” The link takes you to a page that contains 25 to 30 explicit and detailed drawings of children performing sexual acts. In a May 6 2010 discussion, Wales spoke specifically about the above image and others in the “pedophilia” and “zoophilia” categories (the latter includes illustrations of children engaged in sex acts with animals)."

  • Jimmy waffled a bit which, given circumstances and time frame, doesn't seem to detract from taking moral high ground. First, Jimmy said he would delete the Nude Children Commons category, then reversed his opinion to keep. 4 hours later, he decided to delete "for lack of educational value". A law professor consulted by Fox said, "With respect to child pornography, the real harm is in the production of it -- not the fact that it's also socially irredeemable." A lawyer who prosecutes child abuse cases disagreed.
  • Did the FBI pursued a case against WMF?
  • Were the child bestiality and pederasty drawings deleted from Commons or just from Wiki articles?
  • Why did The Signpost omit Jimbo's focus on the removal of child porn and child bestiality content from WMF servers? He wasn't merely expunging artworks depicting nude adults or naked youth in non-sexual contexts.--FeralOink (talk) 23:12, 20 February 2023 (UTC)[reply]
There is a very good reason: As far as I can tell None of the works Jimbo deleted in 2010 had anything to do with children or animals.. Here is a list of every file Jimbo deleted. If you can spot a single child or bestiality image in there... Adam Cuerden (talk)Has about 8.2% of all FPs. Currently celebrating his 600th FP! 23:35, 20 February 2023 (UTC)[reply]
@FeralOink: Basically, there's literally zero evidence he deleted a single image on grounds that it was illegal or immoral. I don't think we have the works Sanger commented on anymore - but that has nothing to do with Wales' deletion spree. It's hard to check now - and I'm not sure I'd want to - but I believe all such images were deleted long before Wales began his attack on adult sexuality. We can argue about whether what he did delete should have been, but as far as I can tell it falls into the following categories:
  1. Artworks, mainly by van Bayros, who, while explicit, is showing consentual adults.
  2. Illustrations of sexual acts for articles on them, all of which are for adults
  3. Explicit pictures of adults
  4. Adult pornographic content.
It may be that some of it - especially from the latter two categories - would have been deleted as out of scope in a deletion discussion. But there's literally nothing in the list that I can see that could justify the extreme breaking of procedures.
Sanger's complaint may have been valid - I can't and don't want to see what was deleted in response to his complaint - but it doesn't have the slightest bit to do with Wales's actions. Adam Cuerden (talk)Has about 8.2% of all FPs. Currently celebrating his 600th FP! 23:39, 20 February 2023 (UTC)[reply]
You have no idea whether or not Wales deleted any child porn, because child pornography images are expunged from Wikimedia servers; it's illegal to store them in any way. Deleted and suppressed files/images/articles/other pages remain on the servers, in case you weren't aware, even if they aren't viewable by the general public; therefore, when expunged, it is likely that any record of the existence of those "documents" is also expunged. That would include deletion, undeletion and suppression logs, as I understand. It is my recollection that there were indeed quite a few images that could have been perceived as child porn that were involved in that deletion effort. Risker (talk) 07:37, 21 February 2023 (UTC)[reply]
Possible, but Wales never said anything about it at the time that I can see, and quite a lot about wanting to remove sexuality in general. I can't prove it, but as I remember, people were pointing out at the time that he was utterly ignoring anything related to that but mass deleting adult sexual content and artworks. Adam Cuerden (talk)Has about 8.2% of all FPs. Currently celebrating his 600th FP! 15:11, 21 February 2023 (UTC)[reply]
Commons still hosts explicit drawings by Martin van Maële depicting child sexual abuse and bestiality today. There is even a well-stocked Commons category "Erotic activities involving children", and some of the drawings are fairly widely used in Wikipedia mainspace. I guess WMF lawyers must have decided they can be legally hosted. Andreas JN466 13:37, 21 February 2023 (UTC)[reply]
  • Wikipedia community speaks for itself I have personal involvement in the above case. I also have involvement through Wikimedia LGBT+, which is a organizational stakeholder in the case due to holding scope in matters of sexuality and gender. When this happened both the media and the Wikimedia Foundation routinely and without hesitation disparaged the Wikimedia community often, and the Wikimedia community was less able to defend itself from attacks or answer accusations. I do not feel that this case was about inappropriate content in Wikipedia; instead it was an attack on the Wikimedia's community's excellent and working moderation and review practices, which were successfully preventing the hosting of inappropriate content. When I think of this case, I think the Wikimedia community protected the ethics and values of the Wikimedia movement, and other players from outside the community, including in the Wikimedia Foundation, engaged in misconduct. When there is interest from journalists or researchers the Wikimedia community still has its own perspective to share and stories to tell. The Wikimedia community also has standing to demand that the Wikimedia Foundation clarify their positions, such to disclose in an appropriate way if they actually found inappropriate content. If they did, there is no public record. It is inappropriate to fail to provide records of what happened because that sustains the accusation that the Wikimedia community was allowing inappropriate content, when we have no evidence that this happened.
Here is my summary based on the public records: Hatemongering groups made false claims that Wikipedia was hosting inappropriate content. The Wikimedia Foundation through its representatives deleted certain content, and their process included making accusations and attacks against content which Wikimedia community consensus has deemed appropriate for hosting in the platform, including media for the arts, sexual education, erotica, and pornography. The long-lasting harms from this include lowered reputation of the Wikimedia community in the public imagination, disenfranchisement of Wikimedia community members as participants in our democratic system, censorship to align with arbitrary morals which the community rejects, and false Wikimedia Foundation claims that they as an organization have the right to conflict with and override the Wikimedia community's judgements on matters of ethics and values.
LGBT+ people are attacked from all directions by organizations which attack expressions of sexuality, gender, and freedom of expression. My opinion is that the Wikimedia Foundation's behavior in this case is rightly described as homophobia, and the correct response for any future such situation would be respect for community stakeholders through a community conversation process which the Wikimedia community itself confirmed was without power disparity.
Always say NO to anyone speaking on behalf of minority groups! Let the people speak for themselves! Bluerasberry (talk) 19:21, 21 February 2023 (UTC)[reply]
It is notable that the only non-Wikipedian artworks Jimbo deleted were related to lesbians and female sexuality. But that's often how censorship works: LGBT works get much stricter scrutiny. Adam Cuerden (talk)Has about 8.2% of all FPs. Currently celebrating his 600th FP! 19:50, 21 February 2023 (UTC)[reply]
  • One other consideration needs to be included here: had any other Wikipedian arbitrarily deleted images as Wales had, they'd be blocked, their case taken either to WP:AN/I or the ArbCom, & possibly banned for life. Ignore all rules will only support an action like that so far. I will admit that at the time it was not well known just how utterly wrong Larry Singer is about Wikipedia in practically all of his statements (unlike now), so perhaps Wales' overreaction can be defended to some degree. -- llywrch (talk) 20:59, 21 February 2023 (UTC)[reply]
    I think the big problem is this explicitly wasn't about Sanger's complaint. Wales' statement was "I think a perfectly legitimate position for us to take is that we don't have visual depictions of explicit sexual activity here. I think it's a perfectly fine thing to have people collecting classic pornography – on their own servers, separate from Wikimedia completely."
    That's not even a response to Sanger's comment. That's Wales unilaterally deciding that sexual content *as a whole* is bad. Adam Cuerden (talk)Has about 8.2% of all FPs. Currently celebrating his 600th FP! 22:13, 21 February 2023 (UTC)[reply]
  • Hello Blue Rasberry and Adam and Risker et al. I noted two artworks that Jimmy Wales deleted, and I am not sure why. They didn't seem to be about lesbians or the gay community. They were in the medium of line drawings, not photos or color paintings, with provenances in the late 1800s. They were not the work of "major artists". One seemed pleasant and totally innocuous to me. Maybe they were naked women lounging, I couldn't really tell. Another depicted what appeared to be an intentionally caricaturized depiction of a middle-aged working-class woman of the 1800s, sitting on the ground, inebriated and disheveled, with her mouth opened wide around a large dildo. I do not know why Jimmy wanted to delete these artworks. I do know that he was questioned stridently by the editor community for his deletion of a Commons image of a 14 year old girl having intercourse. English Wikipedians claimed that age of consent was 14 in German, and that Americans were prudish idiots whose provincial morality is determined by Fox News. At least one German Wikipedia editor stated that it was NOT legal in Germany to depict 14 year old girls having intercourse even though it was legal to actually have intercourse at that age. Over the course of several threads, the German guy asked with increasing urgency that such material be deleted by Jimmy. Then there were a whole series of photos by one Commons editor who has since been banned of women being penetrated with objects and beaten by men (according to the photo image file names and the captions) and these images had nothing to indicate they were about consensual BDSM type activity. Some were retained, some deleted. There were also some discussions about what might have been underage males with exposed genitalia and lots of images with file names "piss in ass". There are STILL a lot of young male auto-fellatio videos which might be of minors, I don't know. They aren't used in Wikipedia articles. I watched them all because of my prurient interest, and have never seen anything like them on any Internet porn website. I digress. So, Jimmy deleted about 100 images "for lack of educational value" and several versions of the same two artworks. I can't explain the reason for the latter two and am glad they were restored. I am fond of you, Blue Raspberry, and perhaps you are referring to other activity of which I am unaware. There was no gay community content involved in this incident that I could tell, and Adam hasn't provided any examples of such.--FeralOink (talk) 08:51, 26 February 2023 (UTC)[reply]
@FeralOink: I think each person who looks at the currently available discussion and records will come to different conclusions about what happened. I reiterate that I both see LGBT+ content in the collection, and I perceive the significant and conscious absence of LGBT+ engagement in the issue. Few people want to enter a discussion which major media sources and the Wikimedia Foundation itself have labelled as child exploitation.
This comment section is not the place to sort it. At any time in the future, if and when Wikipedia is of interest to journalists or researchers in media studies or gender studies, then I think there is more to tell about this story even with contemporary examination of the available records. The story that I would want to tell is 1) the wiki community sincerely and effectively creates and upholds ethical standards and 2) it is not the place of external powers to swoop in, claim authority, and circumvent the community's ethical review process. The wiki community way is transparency and public discussion; it was the WMF choice in this case to avoid that. Transparency and frank discussion would still make this occurrence better. Bluerasberry (talk) 13:17, 27 February 2023 (UTC)[reply]
  • I don't know how to format things here, what your conventions for threading are. I apologize for making a mess. Adam Cuerden, you keep saying that Larry Sanger wasn't the catalyst for Jimbo's actions. But he was!
  1. Sometime in March or April 2010, Larry Sanger made a formal complaint to the FBI, about child porn on Wikimedia servers that was then included in Wikipedia articles.
  2. On 27 April 2010, Fox News published a news report about it.
  3. On 7 May 2010, Fox reported that Jimmy Wales and Wikimedia officials were assessing and deleting pornographic images of children from WMF servers. I found evidence of Jimmy Wales deleting some adult sex photos, a few instructional images contributed by Flying Lady (I forgot her user name, but she died about a year or two ago), a few drawings of adolescent males masturbating, as well as some photos of adult females being physically beaten that were Wikipedia-editor contributed, and a few naked images of possibly 14 year old to 21 year old women posing on beaches etc. Fox didn't report at that level of detail.
  4. On 10 May 2010, Fox wrote the last of its three part series on the subject. That is not consistent with The Signpost's description of a single Fox News article on 10 May 2010, "attacking Wikimedia".
  • Also, there was vigorous discussion about whether some images Jimbo wanted to delete were of children. I don't have any special user rights. I merely perused the links you provided in your post.
  • I am not a Jimmy Wales apologist! (Not that I am advocating for Larry Sanger either, as, well, it is a sad situation but Illyrch already addressed it.) About the matter of Jimmy Wales giving up founder privileges, I'm not even so sure of that. You might be saying that prior to May 2010, Jimbo could edit Wikipedia with the same impunity as one could edit one's personal Wordpress blog, and now he can no longer do so. In other ways, Jimmy Wales retains plenty of founder privileges, e.g. he is the permanently designated "Founder Community Director of the Wikimedia Endowment Board" with oversight of who knows how many hundreds of millions of dollars if they ever withdraw the money from Tides Foundation. (It blew through its goal of $100 million in assets under management a full 5 years ahead of schedule in May or maybe September 2021.) The point of my objections here are that it is misleading to suggest that Wikipedia has taken back moral high ground that it might or might not have ever had. Also, there are lots of gay Republicans and gay watchers of Fox News and they aren't pedophiles or self-hating. I am going to stop fussing over this now. (I care about Blue Raspberry and hope my statement has not offended him, as I would never want to do that. He is nice to me and also cares about financial and governance propriety and honesty like I do.)--FeralOink (talk) 09:18, 26 February 2023 (UTC)[reply]
Thanks for sharing clarification. 💙 Bluerasberry (talk) 13:22, 27 February 2023 (UTC)[reply]

Disinformation report: The "largest con in corporate history"? (13,960 bytes · 💬)[edit]

Discuss this story

Great reporting. Also interesting in this context is this recent Wired article titled "India’s Government Wants Total Control of the Internet":

[Akash Banerjee is] a seasoned journalist who runs The Deshbhakt (“the patriot”), a satirical YouTube channel covering politics and international affairs.

[...]

[Adani Group] has denied the allegations and responded using nationalistic language, calling the [Hindenburg] report a “calculated attack” on India and its growth story. Days after the report came out, a video about Adani Group founder Gautam Adani that Banerjee had posted four months before was suddenly targeted by YouTube on the grounds of “strong profanity”—presumably following a user’s complaints.

“I had called him [Adani] an oligarch,” said Banerjee. “But the video was demonetized, saying it has profanity. Is oligarch a profanity? Then I don’t know.”

Although he’s still posting and has yet to face any serious legal problems, Banerjee says he’s already making preparations for that possibility.

“Any social media person, anyone who is willing to do commentary, I always say they have to have two very good things,” he says. “A good chartered accountant and a good lawyer.”

Regards, HaeB (talk) 11:32, 21 February 2023 (UTC)[reply]

Thanks @HaeB: I hadn't seen anything from Banerjee before, but there are still many videos of his on YouTube. I'd say just turn on the subtitles and you can get it in English, rather than Hindi(?), but his presentation, speed, sense of humor and switching between languages makes it very difficult for me to understand in any case. I don't think he is backing off from any attempted censorship though.
Neither are about 7 Indian newspapers (if they are being threatened with censorship) who are covering this article from The Signpost.
  • The Wire (India) Paid Users, Including Employees, Improperly Edited Adani Articles, Says Wikipedia Newspaper (I love the illustration at the top!)
  • Business Today Did Adani's team 'systematically manipulate' Wikipedia entries?
  • Economic Times Wikipedia editors blame billionaire’s team for manipulating entries
  • Deccan Herald Sockpuppets' created puffery about Adani, says Wikipedia
  • Asianet Newsable Paid users, including workers, inadvertently edited articles about Adani: Wikipedia (strangely "inadvertently" only appears in the headline)
  • Mint
  • The Hindu
  • Times of India

You might even say that this incident is notable. Smallbones(smalltalk) 16:17, 21 February 2023 (UTC)[reply]

  • One of the unintended consequences of the capitalization style on Wikipedia is that headings written in title case tend to be a red flag.~TPW 19:41, 21 February 2023 (UTC)[reply]
@Smallbones, that's impressive coverage of your coverage. However, it's now been revealed that "Signpost recently banned over 40 sockpuppets". That's very out-of-process of you (plural), and I will of course bring it to ANI. Gråbergs Gråa Sång (talk) 09:55, 22 February 2023 (UTC)[reply]
Btw, this page needs some Template:Press. Gråbergs Gråa Sång (talk) 10:39, 22 February 2023 (UTC)[reply]
@Gråbergs Gråa Sång: - that's obviously a mistake that the newspaper made - nothing to do with me, I've never even heard of them before. And I hope you're joking about ANI. One thing that somebody can do for me, please tell that newspaper that The Signpost can't ban anybody! Their text "Signpost recently banned over 40 sockpuppets ... Wikipedia reported that all 40 sockpuppet accounts have been banned," should be changed to "Wikipedia administrators banned over 40 sockpuppets ... 'The Signpost reported that all 40 sockpuppet accounts have been banned or blocked." Also the bans and blocks were generally not recent. For a private reason I'd rather not tell the newspaper this myself, and why would they believe that I'm Smallbones. Paradoxically, if you or somebody else tells them about their mistake and to see this page, it would confirm that a mistake has been made better than I could. Thanks for pointing this out. Smallbones(smalltalk) 12:30, 22 February 2023 (UTC)[reply]
I was indeed joking. Amusing misprint, that's all. Gråbergs Gråa Sång (talk) 14:16, 22 February 2023 (UTC)[reply]
Thanks. I see the joke now! Smallbones(smalltalk) 16:49, 22 February 2023 (UTC)[reply]
Sorry if my attempt at humor caused you any distress or annoyance, that was not the intent. Gråbergs Gråa Sång (talk) 19:26, 22 February 2023 (UTC)[reply]

And in video.[edit]

There are now 2 videos on this article. It's rather strange seeing a video in Hindi going over my work almost word by word. It's very clear that the two words they are most interested in are "sockpuppet" and "puffery". See

Hey @Smallbones It bothers me that you say Adani manipulated all the mentioned pages although only one user Satyam Trivedi was unmistakably (by his/her own admission) identified as an Adani staffer in 2013. Pardon my ignorance here but how do we know for a fact that everyone else was working for Adani? Manatpeace (talk) 18:29, 21 February 2023 (UTC)[reply]

sometimes you have to settle for beyond reasonable doubt.©Geni (talk) 20:50, 21 February 2023 (UTC)[reply]
  • Thank you. Got it. Manatpeace (talk) 02:49, 2 March 2023 (UTC)[reply]
  • @Manatpeace: I certainly did not say "for a fact that everyone ... was working for Adani." That would mean that I believe that 40 accounts that had been blocked for socking (after their edits on these articles) had been working for Adani. I don't believe that at all in fact I'd be a quite surprised if at least 1 was not working for Adani! There's an element of uncertainty here for each of the blocked socks. While I haven't tried to quantify the uncertainty, it might be worth the exercise. Let's say, for argument's sake, that each of those blocked socks who edited these articles had a 50% chance of working for Adani or his companies. In that case the probability that not all 40 having worked for Adani is 1- (1-0.50^40) = 0.99999 (for at least 10 digits) in which case we can say something like "they almost certainly were not all working for Adani," or more colloquially "It would be a miracle if they all worked for him." But I didn't say that they all worked for him.
I wrote "So did Adani or his employees improperly edit Wikipedia? Almost certainly yes." The calculation here is for the probability that *at least one* of the blocked socks worked for Adani when each one has a 50% chance of working for him. It's essentially the same calculation 1 - (0.50^40)=.99999 etc. I'll say "It would be a miracle, if at least one didn't work for Adani!"
Well, that wasn't my exact thinking, but the question is "how much evidence do you need until you're sure enough that Adani or his companies were responsible for the socking?" I said to myself "40 seems like quite enough." Several of them, e.g. Hatchens, Similar2me, UncleScrooze, have very high probability of working for the Adanis. Some might only have a 20% chance. Hope that helps. Smallbones(smalltalk) 21:10, 21 February 2023 (UTC)[reply]
I did the math. This is similar to the birthday problem. Starting with a model where the probability that each of the actors meets the criterion "working for A" is independent, and there are 40 actors (blocked socks). In this case you reach 50% probability that at least one meets the criterion at about 1.7% probability for each individual actor 1-e^((ln 0.5)/40) = 0.01718. As the probability of a single actors meeting the criterion increases, obviously the probability that at least one in the group meeting the criterion also increases: e.g. probability of just 5% for each actor results in over 87% odds overall 1-((1-0.05)^40) = 0.8714. And I concur with Smallbones' math above that 50-50 odds for each actor result in 10 or 11 nines, an "almost certainly true" conclusion of some subset of the group acting in cahoots with the subject. ☆ Bri (talk) 17:45, 23 February 2023 (UTC)[reply]
Thank you. Wasn't doubting your judgement anyway! Manatpeace (talk) 02:48, 2 March 2023 (UTC)[reply]

Essay: Machine-written articles: a new challenge for Wikipedia (11,767 bytes · 💬)[edit]

Discuss this story

  • I've found that ChatGPT can be good at writing articles. I asked it to write an article on The Crew Motorfest (out of curiosity, not to actually write the article as I had already been written and published) and the article came out pretty decent with only 1 inaccuracy I found. I asked it source it and it came up with BS sources however so it isn't perfect. ― Blaze WolfTalkBlaze Wolf#6545 18:15, 20 February 2023 (UTC)[reply]
    Also, mind posting a link to the AI detector you used? ― Blaze WolfTalkBlaze Wolf#6545 18:16, 20 February 2023 (UTC)[reply]
The links are in the "Finding More" section, the primary one is https://openai-openai-detector.hf.space/ ☆ Bri (talk) 18:21, 20 February 2023 (UTC)[reply]
Thanks! I didn't see a finding more section but I Could just be dumb. ― Blaze WolfTalkBlaze Wolf#6545 18:23, 20 February 2023 (UTC)[reply]
ChatGPT creates plausible-sounding bullshit. In cases where it has a lot of very similar sources to draw from, such as mostly-empty space-filler articles about an upcoming racing video game (for which it would have about a thousand examples) it can generate something low on nonsense. For something more unique, the bullshit quota is higher. In all cases, though, you can't tell what's bullshit without checking it line by line, because it's all plausible-sounding. Similarly, the sources will always be nonsense, because it isn't generating text based on specific sources, it's generating plausible-sounding reference text bullshit, with no connection to anything. --PresN 19:29, 20 February 2023 (UTC)[reply]
Yes I'm not trying to argue that we should be using ChatGPT (because frankly no one should), simply that it isn't 100% bad all of the time. ― Blaze WolfTalkBlaze Wolf#6545 19:31, 20 February 2023 (UTC)[reply]
IN fact I have encountered situations where it likes to hallucinate (I asked it a few things regarding Splatoon and it kept thinking the special gauge was the amount of ink the weapon had which is not true whatsoever) no matter what I tell it. ― Blaze WolfTalkBlaze Wolf#6545 19:33, 20 February 2023 (UTC)[reply]
One of the data sources for ChatGPT is Wikipedia, so if you ask it to write about something already in Wikipedia, there’s a likelihood that it will select correct information for its output. — rsjaffe 🗣️ 22:24, 20 February 2023 (UTC)[reply]
WP:Randy in Boise can also make good contributions most of the time, but the few times he's wrong still make him a net negative. AI seems to be a long way from getting past this level of ability. Daß Wölf 20:24, 24 February 2023 (UTC)[reply]

I test for articles typically using https://openai-openai-detector.hf.space/ - this and various other currently available "ChatGPT detectors" (including OpenAi's own) are highly unreliable. https://openai-openai-detector.hf.space/ actually already says on the tin that it is a detector for GPT-2 (released in 2019 and very different from ChatGPT). Given the article's focus on the dangers of misinformation, it's a bit sad and ironic that the Signpost is itself providing such dubious recommendations here without any caveats.

Regards, HaeB (talk) 11:12, 21 February 2023 (UTC)[reply]

The article glosses over a lot of the issues regarding detection. It was just a brief intro. I emphasized in the article that I was using a very insensitive method of finding LLM-generated text. There were a couple of reasons I went about things as described there (and to note: I no longer rely solely on GPT-2 detector). 1) at the time I started, other detectors available were very opaque as to how they were constructed; 2) the nature of the output, even though the models are different, has many similar characteristics, so a GPT-2 detector would have some sensitivity and specificity; 3) I intentionally minimized false positives as those irritate article contributors, by doing a vigorous pre-screen of the text. As to point two, note that at least one of the recommended detectors (https://gptzero.me/) is not based on the GPT model, but rather on the text output characteristics. As to point three, I used the authors' feedback as an indicator of the false positive rate: getting no complaints after a lot of tags is a decent indicator that the false positive rate is low. — rsjaffe 🗣️ 18:53, 21 February 2023 (UTC)[reply]
Good to hear that you are proceeding diligently when patrolling new articles (and to be clear, this is very important work and it's good to call attention to this issue). But the part with the tool recommendations was not including any caveats about false positives, and should not have been published in this form.
the nature of the output, even though the models are different, has many similar characteristics, so a GPT-2 detector would have some sensitivity and specificity - what research is this claim based on? (I mean, of course any detection method has "some sensitivity and specificity", the question is whether they are good enough.)
is not based on the GPT model, but rather on the text output characteristics - it seems that there is some fundamental confusion here between the model that is doing the detection and the model whose output is being detected (and/or the features of its output). https://openai-openai-detector.hf.space/ is also not using "the GPT model" (there are many actually) to detect GPT-2 output, but RoBERTa instead.
Regards, HaeB (talk) 06:57, 24 February 2023 (UTC)[reply]
  • I'm not convinced that this is a problem that isn't already addressed by the vast number of policies and guidelines on this site. If artificial intelligence ever becomes capable of generating Wikipedia articles that are verifiable, written in a neutral point of view and devoid of original research, then I'm all for it. Until then, the usual system of separating cruft from quality will continue. It's possible that garbage will be generated faster than ever, but that seems like a technical issue rather than a policy one.~TPW 19:46, 21 February 2023 (UTC)[reply]
    Read the proposed policy. I'd look at the LLM policy more as an explanation as to how text generation fits into current policies rather than setting new precedent. The problem is that most people do not understand the policy issues raised by LLMs. The proposed policy explains them. — rsjaffe 🗣️ 20:29, 21 February 2023 (UTC)[reply]
  • If AI is integrated into Wikipedia to write articles and whatnot, we should have some sort of Pending Changes Protection when AI is used so we could double-check the accuracy of the article(s). ‍ ‍ Helloheart ‍ ‍ 00:49, 22 February 2023 (UTC)[reply]
    If robots can reliably write articles of ordinary WP quality, then there's no need for WP. People who want to know something can just ask the robot and get an answer tailored to the asker's known preferences and knowledge. Jim.henderson (talk) 20:46, 22 February 2023 (UTC)[reply]
    Wouldn't that make for a bubble, though? Here's an article to prove whatever you already believe! Adam Cuerden (talk)Has about 8.2% of all FPs. Currently celebrating his 600th FP! 03:14, 23 February 2023 (UTC)[reply]
    How do I know I'm smart? My computer friend always tells me I'm right, that's how. Tyrants have suffered bad advice from ego-stroking yes-men forever; now everyone can be Ethylred the Unready.Jim.henderson (talk) 02:49, 26 February 2023 (UTC)[reply]
  • Hallucination is a major problem with these models, anyone not verifying each and everything they are using AI tools for should be immediately sanctioned. There was an interesting discussion to that effect on Villagepump I believe, even before these tools came into vogue. Gotitbro (talk) 13:10, 1 March 2023 (UTC)[reply]
  • Good and timely article. I also tried out ChatGPT to see what it can do. It's documented on this page. My overall conclusion was that "it appears that us Wikipedia volunteers aren't out of a job just yet". Schwede66 22:23, 7 March 2023 (UTC)[reply]
  • After reading another article about how ChatGPT is coming for all our jobs, I signed up and asked it (3.5 I presume) "how many neutrons are in a liter of water" (yes, it insisted on that spelling of litre). It wrote out a four or five paragraph reply explaining exactly how it arrived at the figure of 556 neutrons. Typing the same into the integrated engine in Bing resulted in the claim that a litre of water has no neutrons. 4.0 did somewhat better, off by only two orders of magnitude. So I think the bit about "AI-generated text is not reliably correct" needs to be bolded. Maury Markowitz (talk) 20:34, 18 March 2023 (UTC)[reply]

Featured content: Eden, lost. (0 bytes · 💬)[edit]

Wikipedia talk:Wikipedia Signpost/2023-02-20/Featured content

From the archives: 5, 10, and 15 years ago: Let's (not) delete the Main Page! (946 bytes · 💬)[edit]

Discuss this story

For more events in enwiki history that are similar to the "don't delete the main page" shenanigans, see Wikipedia:Village stocks. I was introduced to it forever ago by someone telling me it was full of examples of what not to do. It's been around forever but it's worth a read to newer wikigenerations who maybe haven't seen it? Clovermoss🍀 (talk) 07:28, 22 February 2023 (UTC)[reply]

Wow. We deleted the main page fifteen years ago. Time flies. I remember being there for that one big time. Wish I still had the irc logs or whatever it was we were all goofing off in back then. Hiding T 12:48, 7 April 2023 (UTC)[reply]

Gallery: Love is in the air (0 bytes · 💬)[edit]

Wikipedia talk:Wikipedia Signpost/2023-02-20/Gallery

Humour: The RfA Candidate's Song (2,452 bytes · 💬)[edit]

Discuss this story

  • A great parody, and fun to sing along. Incidentally, was there ever really a time when someone would oppose an RfA because the candidate had not participated enough in portal talk, or is that just a joke? In 2023 I think you could become an admin without ever having seen a portal. CharredShorthand.talk; 18:44, 20 February 2023 (UTC)[reply]
  • @CharredShorthand: There was a period where checking what namespaces you participated in was a huge part of Adminship. Something something well-rounded something something understands all aspects. So more exaggeration. Adam Cuerden (talk)Has about 8.2% of all FPs. Currently celebrating his 600th FP! 22:00, 20 February 2023 (UTC)[reply]
What's a portal? (/s) Lee Vilenski (talk • contribs) 19:27, 20 February 2023 (UTC)[reply]
  • I remember singing this back in middle school. Wow, 17 years later and it's just as relevant. -- RockstoneSend me a message! 23:17, 20 February 2023 (UTC)[reply]
  • We need more hilariously stupid stuff like this. Unfortunately various modern policies and essays and stuff sort of discourage/prevent this. ― Blaze WolfTalkBlaze Wolf#6545 00:41, 22 February 2023 (UTC)[reply]
  • Like I said in January in this publication, "fun 'is no longer relevant or consensus or its purpose has become unclear'". ☆ Bri (talk) 00:53, 22 February 2023 (UTC)[reply]

In the media: Arbitrators open case after article alleges Wikipedia "intentionally distorts" Holocaust coverage (6,865 bytes · 💬)[edit]

Holocaust arbcom case[edit]

Just a note for precision that the heading "Arbitrators open case ..." seems technically speaking a bit incorrect, since ArbCom was only "initiating a case request" (as the article body correctly says). However, given that the arbitrator votes currently stand at 9/1/0 (accept/decline/recuse), it can be safely assumed that the case will in fact be opened. Regards, HaeB (talk) 11:51, 21 February 2023 (UTC)[reply]

  • One thing that the authors of the paper got right, but the authors of this article got wrong, is who is being accused. The paper's authors refer to a group of editors, while the article's author suggests that Wikipedia itself is responsible. The agency always rests with the people. That shorthand evades responsibility in a way that is common among government and corporate officials, and doesn't belong on this site. Wikipedia is not paper, and there's no reason not to be precise.~TPW 14:09, 21 February 2023 (UTC)[reply]
    As the author of the article in question, I'd like to point out that Grabowski and Klein titled their essay "Wikipedia's Intentional Distortion of the History of the Holocaust".
    Grabowski and Klein did refer to editors in the abstract—which I quoted in full. (I wrote neither the main headline nor the sub-headline for this page, but they don't seem to me at variance with the title of the essay.) Andreas JN466 17:19, 21 February 2023 (UTC)[reply]
    • From Grabowski's 2020 Polish newspaper article, something that may help understand what he thinks about our project in general: [1] I warn students against using Wikipedia and I advise the use of paper encyclopedias. ... I hesitate to use the term 'regulation', but it will be difficult to win otherwise with the spread of lies. Perhaps the time has come for supranational, pan-European institutions... to take a closer look at Wikipedia. Fake news is doing great there... . I just dusted off the 20 volumes of Meyer's encyclopedia that I inherited from my grandfather and had been sitting unused in the family library for years. Using it, I can be sure that articles will not change beyond recognition tomorrow. --Piotr Konieczny aka Prokonsul Piotrus| reply here 12:08, 22 February 2023 (UTC)[reply]
      I remember looking in an encyclopedia my grandparents had. The entry on Adolf Hitler was something like "German politician. Attempted a coup in 1923, sentenced to 5 years in prison."Gråbergs Gråa Sång (talk) 06:42, 23 February 2023 (UTC)[reply]
      I wonder which edition of the Meyers Konversations-Lexikon his grandfather had... Piotr Konieczny aka Prokonsul Piotrus| reply here 06:54, 23 February 2023 (UTC)[reply]

Updated link[edit]

The link to Volunteer Marek's response was updated on 21 February, 2023. It now leads to a different target. (The other document is also still up and linked from there.) --Andreas JN466 16:51, 21 February 2023 (UTC)[reply]

Baker[edit]

Baker is correct on the negative bias in biographies of even slightly controversial figures either due to media coverage as such or the tendency of editors to focus on them. The way to mitigate that is ready access to non-news sources which cover the subject in toto. I wonder what he has to say about FA articles (though these tend to be few for controversial topics). Gotitbro (talk) 13:41, 1 March 2023 (UTC)[reply]

I have opened a discussion at Wikipedia:Neutral point of view/Noticeboard#Russ Baker to address these complaints. -Location (talk) 16:59, 7 March 2023 (UTC)[reply]
I was involved in that article a while back, and it's one of the reasons I've tended to avoid COI-laden subject areas, paid editing and so on, which used to be a major focus, I don't follow the COI noticeboard anymore and have given up on paid editing. The COI editor, whether the article subject or a PR firm, is paid or otherwise inherently prone to never give up, not give a damn about Wikipedia policies, and hammer away until they get their way. I started on Wikipedia because of BP, where a PR man was working full time to slant an article. I don't edit BP anymore because dealing with such a determined article subject is simply not worth my time. I think COI situations like BP and this one are a major reason for editor burnout. One reason for this is that other editors very often couldn't care less about COI situations.
What bothers me about the Signpost article is that it fails to point out that the subject of the article actively participated on the talk page and in the article itself. Dealing with COIs can be very stressful, I've certainly encountered a lot worse. One of the subjects of a Trump pardon dispatched an army of sockpuppets to the page on the pardon subject, who was released from a very long prison term. They wanted the article to be a love song, You can't blame them but, please, show a little empathy for the volunteers who have to deal with people like that, who have absolutely no interest in Wikipedia policies. Coretheapple (talk) 19:44, 10 March 2023 (UTC)[reply]
Just to quantify one point I made above. Baker has been editing this article through SPAs and IPs for at least thirteen years. See [2]. Coretheapple (talk) 14:57, 12 March 2023 (UTC)[reply]

Judges[edit]

Senior judges using Wikipedia is not surprising but I wonder what lead to such a situation, easy accessibility perhaps? Gotitbro (talk) 13:44, 1 March 2023 (UTC)[reply]

News and notes: Terms of Use update, Steward elections, and Wikipedia back in Pakistan (5,393 bytes · 💬)[edit]

Discuss this story

How can the license for text be changed to the Creative Commons BY-SA 4.0 license retrospectively? Doesn't the old license apply in perpetuity? If not, shouldn't we take the opportunity to change it to a CC-NC license? Hawkeye7 (discuss) 23:47, 20 February 2023 (UTC)[reply]

  • @Hawkeye7: as I understand it, the reason for changing the ToU now is that several small changes to it have been proposed and tentatively accepted, but that going thru the ToU process is daunting enough that they decided to wait until something big came along to make one big change rather than several little ones. The two big things now are UCoC and the new European law. The tentative decision for changing to CC BY-SA 4.0 was made all the way back in 2016(!) by an !vote of 319-92 in favor. This is at m:Talk:Terms of use/Creative Commons 4.0 where the FAQ says:
How can we "upgrade" the license?
If we choose to amend the Terms of Use, the 4.0 version of the license will apply to new edits submitted to Wikimedia projects. After a page has been edited, it can be reused under the latest version of the license according to the attribution requirements in the Terms of Use. Revisions of pages before the upgrade to the 4.0 version will continue to be available under the version 3.0 of the license.

I love that you asked ChatGPT for a statement, and it's AI-generated opinion seems pretty spot-on! WaggersTALK 09:42, 21 February 2023 (UTC)[reply]

Tips and tricks: All about writing at DYK (1,178 bytes · 💬)[edit]

Discuss this story

  • @Theleekycauldron: As an user who is trying his hand at DYK, both as a nominator and a reviewer, this essay is a gold-standard source of help. Thank you so much for it! For example, I think my hook for Valentín Carboni would likely fall in the "yellow" area, rather than in the "green" one, wouldn't it? That would explain its average performance... Apart from jokes, I hope we'll be able to team up together in the future! : D Oltrepier (talk) 11:02, 21 February 2023 (UTC)[reply]
    Thank you so much, Oltrepier! I'm glad someone found a use for it :) you let me know if there's anything I can do for you! theleekycauldron (talk • contribs) (she/her) 06:04, 23 February 2023 (UTC)[reply]

Traffic report: Superbowl? Pfft. Give me some Bollywood! Yours sincerely, the world (989 bytes · 💬)[edit]

Discuss this story

  • Sorry in advance if my message goes off-topic, but I just wanted to address that I'd appreciate some help with improving Bacharach's page, and especially the "Discography" section, so we could give him proper justice. I've already tried my best at tidying up the article and leaving (current) red-links that could serve as useful separate pages, but it's definitely not a task that I can face alone, especially due to my lack of experience... Anyway, thank you for this series! Oltrepier (talk) 11:20, 21 February 2023 (UTC)[reply]


Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia_talk:Wikipedia_Signpost/Single/2023-02-20&oldid=1140569792"