As the dog days, the hottest time of the northern hemisphere's summer, are upon us, there are many solid news stories to report involving Wikipedia, such as the one about a candidate to become the UK prime minister. He dropped out after two days when accusations about his Wikipedia editing were resurfaced. But other stories might be attributed to silly season, the time of year when journalists create odd stories just to fill up space when there's little real news to cover. Dog days and silly season often occur about the same time, but dog days have traditional dates that amateur astronomers generally use, while silly season is a movable feast. Please let us know in the comments section which of the stories you think are part of silly season, and which are real news stories. Feel free to comment on the illustrations as well. – Sb
The Irish Times reports that Research suggesting Wikipedia may influence legal reasoning of Irish judges 'clearly wrong' mainly citing one unnamed source at the High Court. Their complaint is against the conclusions of a chapter to be published in The Cambridge Handbook of Experimental Jurisprudence (Cambridge University Press, 2022), titled "Trial by Internet: A Randomized Field Experiment on Wikipedia’s Influence on Judges’ Legal Reasoning".
The offended dignity of the High Court judges, as conveyed by the High Court anon, might seem to miss the mark. Using a classic experimental design, the first conclusion in the chapter merely shows that the presence of a Wikipedia article on an Irish case increases the frequency of that case being cited later. This would by itself say nothing about whether the judges read Wikipedia to write their opinions. The mechanism of the Wikipedia article's influence could be, as suggested by the anon, that the lawyers read Wikipedia and incorporate the good material into their briefs, which the judges properly study. Other indirect mechanisms of influence are also possible, for example, via interns, clerks, law school lectures, or just a generally more legally literate public.
But the chapter's authors take their analysis a step further. Via a linguistic analysis they test whether the wording of the judges is similar to the wording of the Wikipedia articles. They find that it is, even after removing phrases that quote or paraphrase the cited opinion. High Court judges, you may now properly consider your dignity to be offended. This section on linguistic analysis – rather than the classic randomized control trial – is the key to the chapter.
The authors' interpretations and many reactions to the chapter focus on the perceived threat that an unethical lawyer could bias a Wikipedia article to favor their clients. Yes, that is a theoretical possibility in the future, but there is no evidence in the chapter that this has already happened, and the article writers for the study itself were mostly law students, presumably not yet ruined by practical legal ethics.
The study may have other issues. The authors did the equivalent of search engine optimization by adding infoboxes and short description to the Wikipedia articles, to make sure they appeared near the top of search results. So does this show that Wikipedia really has a strong influence on justice, or simply demonstrate that cases that show up when you search for terms related to them are more likely to be cited?
In any case, it's not unprecedented for judges to cite Wikipedia in judicial opinions: in August 2012, The Signpost reported on how U.S. courts praised Wikipedia, and referenced back to citations of Wikipedia in the England and Wales High Court and other U.K. courts, U.S. courts, and even the European Court of Human Rights and a German patent tribunal. – Sb, AC
National Public Radio, and many other outlets which run the gamut of reliability, report on how the Wikipedia recession article has been semi-protected over an argument about the definition of "recession." Note that the NPR story was significantly modified on July 30, including an explanation from . The argument began when the Commerce Department reported that the US GDP has gone down for two quarters in a row. This measure is a traditional indicator used in part by the NBER to define "recession". U.S. President Joe Biden and other administration officials have said that the current situation is not typical of a recession. A few Wiki editors disagreed and tried, against stiff opposition, to refine the definition. Meanwhile pageviews for the article rose to over 200,000 in one day, perhaps because of the possibility that the U.S. is in a recession, or perhaps because our readers expect Wikipedia to be the arbiter of the media tempest over the definition of recession. As usual the NBER is expected to take their time in making their pronouncement.
Meduza, an independent Russian news site that's banned in Russia but publishes outside of Russia, interviewed Wikimedia RU director Stanislav Kozlovsky about Russian censorship. His surprising response is "When someone points a gun at you for ten years straight, you get used to it."
Kozlovsky points out several facts that may surprise western Wikipedians:
This month Roskomnadzor announced that Russian search engines must put a warning label on Wikipedia articles. Kozlovsky said "The statement from Roskomnadzor’s press service is nothing but a fantasy from their press secretary, and I don’t think it’s going to have any effect on user demand [for Wikipedia]."
If Wikipedia were blocked in Russia, readers could still access the site from mirror sites or virtual private networks. Editing from within Russia might be cut off or slowed, but that would leave the site open to Ukrainian editors. The Signpost will add that many other editors from former Soviet republics, Israel, and the Russian diaspora in Europe and North America could also edit. In short, Russia has nothing to gain from blocking Wikipedia.
"The pressure on us right now is just another charade from Roskomnadzor," according to Kozlovsky. "I also think Wikipedia provides more value to Russia than the Science Ministry, the Education Ministry, and the Culture Ministry put together." – S
Writing in The Boston Globe, Shaun Cammack argues that Wikipedia is in many ways not much different from social media. "Wikipedia is a social media platform where user-generated opinions and interpretations battle it out in a game of popularity and attrition – just like Twitter," he writes. What makes it different than Twitter? Per Cammack, "Conflicts on Twitter proceed as though there were a single record that logs the conclusions of our collective debate. We like to imagine Twitter as a floating metaphysical ledger waiting for that final word. That’s why we dunk on people, fact-check, and correct the record: We think we’re in a reality war and there can be only one winner. But the imperative 'We all have to get on the same page' is literal on Wikipedia. It manifests that imagined ledger as a visible page on which the winning opinion is made concrete and true. And so it might be the most real and significant social media platform of all." – M
Sixth Tone, an English-language online magazine published by the Shanghai United Media Group, reported on 28 June that "A Chinese woman created over 200 fictional articles on Chinese Wikipedia, writing millions of words of imagined history that went unnoticed for more than 10 years." The hoaxer, User:折毛 (transliterated "Zhemao"), was dubbed "China's Borges" online, based on the intricacy of her work, which involved faked sourcing and the use of a number of sockpuppets. The story was subsequently picked up by outlets including the Literary Hub (June 28), Boing Boing (July 7), Vice (July 13) and Engadget (July 14).
Much like the hoaxer reponsible for the Bicholim conflict article (see previous Signpost coverage), Zhemao successfully nominated one of her articles, on Deportation of Chinese in the Soviet Union, for a quality assessment, winning FA status in the Chinese Wikipedia. As a result, a translation of her article was included in the English, Arabic, Russian, Ukrainian and Romanian Wikipedias. The English and Russian versions were substantially revised once the hoax became public; the other language versions have seen no substantial editing to date.
For internal discussions see:
Zhemao posted an apology in Chinese on her English user page explaining how her actions, initially innocuous, grew out of control. The event is now the subject of a dedicated Wikipedia article, which contains further detail: Zhemao hoaxes. – AK
The San Francisco Examiner published a long interview with WMF CEO Maryana Iskander. The 14 questions seem very basic, for example, "What’s the Wikimedia Foundation?", and "Isn’t Jimmy Wales the head of Wikipedia?" Her answers are short and direct and to a very large extent could be used for a great introduction for newbies to Wikipedia. But experienced editors might quibble with some of her answers.
"So how many people get paid and are professional editors of Wikipedia?" Iskander presents the Wikipedia ideal of every editor being an unpaid volunteer. It would be nice, when asked a direct question about paid editors, if she would address our terrible problem with paid editors who use Wikipedia for advertising or otherwise bias our content. This is a major question about the encyclopedia's credibility. Perhaps such a simple interview is not the best place to address this problem, but Wikipedia editors deserve an answer somewhere.
Similarly difficult issues, such as the overall gender gap, or the lack of non-white editors in the US are not mentioned. In these cases the interviewer did not ask the questions, but even so Iskander could have brought up major challenges herself. Wikipedia did not become a great encyclopedia by avoiding tough issues. – S
Facebook AI Research has been organizing Wikipedia and fact-checking evaluation through its Side Editor tool since registering their research on Meta-wiki in January 2022. Now in July, Facebook presents an update and media including TechCrunch, CNBC, and Engadget report that their research may improve both fact-checking in Wikipedia and the broader media ecosystem.
To quote TechCrunch, Meta's AI knowledge tool is called "Sphere" and is –
built around the concept of tapping the vast repository of information on the open web to provide a knowledge base for AI and other systems to work. Sphere's first application, Meta says, is Wikipedia, where it's being used in a production phase (not live entries) to automatically scan entries and identify when citations in its entries are strongly or weakly supported.
In the Wikipedia community Facebook group Wikipedia Weekly, a Wikimedia staffer on the Partnerships Team commented that the Wikimedia Foundation and Facebook have had no collaboration on this project, and communicated that she had Facebook update their documentation to reflect this as can been seen in Internet Archive's website update comparison tool.
Wikimedia and Facebook parent Meta do have an acknowledged partnership when it comes to translation. As communicated by Meta AI on Twitter, Wikipedia editors are using their technology to translate texts into underserved languages. A dedicated page on ai.facebook.com claims that it is "translating Wikipedia for everyone" with "No Language Left behind (NLLB)" and that Wikipedia editors are "using the technology to more efficiently translate and edit articles originating in other under-represented languages, such as Luganda and Icelandic." This Meta AI/Wikimedia partnership was reported this month by MediaPost and The Register.
The NLLB-200 model is open source, and there has been an NLLB-200 page on MediaWiki since January 2022, originally referring to the project under its earlier name of "Flores". A Twitter thread from a long-time Wikimedian argues that
This collaboration is exciting, helpful and problematic – all at the same time.
In the Wikimedia community it is customary that stakeholder engagement, consent, and public conversation about values are part of feature development. If such Wikimedia community engagement exists for these Facebook projects, then The Signpost has not identified where those conversations happened, and we call on readers to identify and document discussions and community collaborations to date. – B, AK
Grant Shapps, the UK's transport secretary, might seem to be the perfect successor to the fuzzy-headed UK Prime Minister Boris Johnson, but he soon dropped out of the race after his past performances were replayed in the press. Australian Broadcasting Corporation wrote how Shapps used the pseudonym Michael Green while selling an online guide which "claimed customers could make $20,000 in 20 days". He's also been accused of deleting embarrassing material from the Wikipedia article about him and adding unflattering material about his political rivals.
The Express gave further details on Shapps's alleged Wikipedia editing in "Grant Shapps humiliation: Minister edited school grades on own Wikipedia page" and reported that he had "pulled out of the race" for prime minister.
The race for the prime minister's office is now down to two candidates. Either Liz Truss or Rishi Sunak will be elected by Conservative Party members by mail-in or online ballots, with results to be reported in early September.
Tom Bower's new book. Revenge: Meghan, Harry and the war between the Windsors, has revived discussion of an old accusation that Meghan Markle, now the Duchess of Sussex, had the Wikipedia article about her changed three weeks before the announcement of her engagement to Prince Harry. The accusations, first from The Telegraph (paywalled), then from Page Six, linked to two edits on October 9, 2016 which eliminated material which made Markle look like a minor Hollywood starlet, and added or emphasized material on the public service aspect of her career that might make her more acceptable to the British royal family.
Page Six reported in 2020 that the then-actress [ Megan Markle's ] Wikipedia page had made the object of a bunch of changes in October 2016, before her months-old relationship with [Prince] Harry was revealed just weeks later. The outlet claimed that the IP address responsible for these changes was "linked to a Los Angeles PR firm with no known links to the duchess."
A Signpost investigation confirms the description of the October 9 edits and also notes that the editor's IP address currently geolocates to Shelter Public Relations in Los Angeles. They have not responded to a request for comment. Tom Bower however told The Signpost that he did not know whether Markle paid the PR company, or whether friends of hers were responsible for the edits, as The Telegram had speculated. More generally, he knows that many celebrities closely watch the articles about themselves. He even offered to name one of his acquaintances who closely monitored his article and removed offensive material. – B, Sb
Complaints that Articles for Deletion is rigged against article subjects are only slightly more novel to most Wikipedians than indoor plumbing, but it's worth reading music historian Ted Gioia's blog post, "How a Prominent Composer Lost His Wikipedia Page—and Got Entangled in Kafkaesque Nightmare Trying to Get it Back," if only to better understand how the deletion process comes across to outsiders.
Gioia makes a case for a page for Bruce Faulconer, best known for his music for Dragon Ball Z, whose page was redirected to that title following a straightforward AfD a few weeks ago that failed to find evidence of significant media coverage or other notability qualifiers. Gioia doesn't find any either, nor does he demonstrate understanding of notability on Wikipedia, instead citing "commissions from dozens of prominent institutions and individuals." From there, he asserts that "petty tyrants connected with the site have imposed their spurious opinions via the platform," citing the user-generated Q&A site Quora.
Gioia says he was referred to deletion review, and he raises the reasonable point that the instructions for listing an entry there leave a lot to be desired, although he mistakenly interprets the complexity as deliberate, writing "This kind of system is clearly designed to intimidate and bully. It's even worse than trying to cancel cable service."
His last-ditch thought to contact the Wikimedia Foundation is preempted by a friend who's already heard that they have no control over content, which he similarly misinterprets as obfuscation.
As the author of 11 books, Gioia is presumably a reliable source on musicians. But if he wants Faulconer to have an article, he should probably write something about his music rather than about his deleted page.