 I am Yumi. I am a Japanese writer based in the U.S. And today I'm going to talk about a disinformation problem on Japanese Wikipedia and how to fix it. Many of you may know that Wikipedia isn't the most reliable source on the Internet, but Wikipedia's misinformation problem runs even deeper than you may realize. So a few months ago, the Wikimedia Foundation which operates Wikipedia released a report called Croatian Wikipedia Disinformation Assessment. In the report summary, the Foundation acknowledged that a group of administrators have been abusing their powers and distorting articles. And they have been whitewashing crimes of World War II and to promote a fascist worldview. And this has been going on for over a decade. And what's interesting about this is that I have found a similar pattern on the Japanese Wikipedia. And I wrote an article about that for Slate. Politically motivated users, editors, have been abusing their powers and distorting articles very much like the Croatian Wikipedia. And they have been whitewashing war crimes committed by the Japanese military during World War II. So I'll give you a few examples. The left shows the page about the Nanjin massacre on the English Wikipedia. And the right shows the same topic on the Japanese Wikipedia. And you can't really see clearly, but the title for the Japanese Wikipedia is actually Nanjin Jiken, meaning Nanjin Incident instead of Nanjin Massacre. And there's no picture of the massacre or events leading up to it, even though such pictures exist, as you can see on the English Wikipedia. There are some pictures. And there's also no highlight of the event. So usually there is an info box. As you can see on the English Wikipedia, there's an info box at the right side. And there is no such thing on the Japanese Wikipedia. Another example is a page about comfort women. There are mostly Korean women who were forced into sexual slavery by the Japanese military. The page uses the word Baishun, meaning prostitution, to describe them, implying that they were not forced. In one passage it says there were three types of battlefield sex policies during World War II, free love, British and U.S., comfort station, Japan, Germany, France, and rape, Soviet Union and Korea. This is not only false, but also incomprehensible. And that seems to be the point. The article is not really designed to inform the readers, but to confuse them and cast a seed of doubt in their minds. And this is kind of how propaganda tends to work. It makes you doubt things that you thought you knew rather than trying to convince you otherwise. So page after page I found examples of historical revisionism. You may think it's just a few pages viewed by some people, but it's actually Japanese Wikipedia is very popular. It's the most visited language in Wikipedia after the English Wikipedia. On average, it receives one billion page views per month. And much like with English search results, every time you search any topic, including historical topics in Japanese, the first page that comes up at the top of search result is usually Wikipedia page. So it's very influential. So each Wikipedia has a talk section that shows the history of discussions editors have had about the article. On the talk pages of controversial historical topics, I found the records of people repeatedly raising concerns and making numerous and successful attempts to correct the content. So there are plenty of people who are concerned about the content and try to change it, but they cannot change it. Historical revisionists in the community often evoke neutrality, one of the pillars of Wikipedia to defend their positions, claiming that their views are neutral. Other times they employ rules such as copyright issues to delete edits and block accounts belonging to users who disagree with them. In short, a group of power editors are controlling the platform. And why is this happening? There are two main reasons, language and culture. So English is a lingaflanca. It is recognized as an official language in a total of 67 different countries. It's spoken by 1.5 billion people around the world. So it's no surprise that English Wikipedia is the largest Wikipedia of all language versions, and currently it has about 120,000 editors. And thanks to the language status as a lingaflanca, editors come from a diversity of background, geographic and cultural backgrounds. In contrast, Japanese is spoken mainly by Japanese. This means that people who edit, administer and read are mostly Japanese, so there isn't as much diversity on the Japanese Wikipedia. Currently Wikipedia exists in more than 300 languages, 302 I think, and half of those Wikipedia have less than 10 editors. So again, English has 120,000 editors, and about 150 language versions have less than 10 editors. So it's clear that many non-English versions are small and lack diversity of editors. And this explains, at least partially, reasons why they have a problem with disinformation. Cultural differences also play a major part in the way technology platforms are used. In a recent interview, Wikipedia's co-founder Jimmy Wales said that Wikipedia was made as an enlightenment project. Some of the important values of enlightenment are freedom of speech, using reason to arrive at truth, and freely discussing societal problems in the public sphere. And these are good values, but they're not necessarily practiced in the rest of the world. So take the concept of the public sphere, for instance. In the West, people have been coming together to discuss societal problems and ideas for hundreds of years. But in Japan, people don't discuss controversial subjects, societal problems in public. People may do that on Twitter or Wikipedia because they're anonymous. But face-to-face, it's very rare for us to engage in those kinds of conversations, and it's because we were told not to actually do those things for the sake of social harmony. So there's not really a concept of public sphere in Japan. So these cultural differences are reflected in the way people use Wikipedia. The talk pages on Japanese Wikipedia show how a group of editors often silence those with opposing views, users who challenge them, risk being accused of political activism, or violating rules, and have their accounts blocked. And it is very similar to Ijime. It's a Japanese term for bullying. It's very common in Japan, and it's a social problem. When I look at talk pages of Japanese Wikipedia, a lot of times this is what I find, a lot of bullying. And it's not like a public sphere. It's far from it. Here are some of other characteristics of Japanese Wikipedia. Over 40% of editors are unregistered and anonymous. As you can see on this graph, it is very high compared to other language versions. And this makes the community more like an anonymous message board rather than the public sphere that Wikipedia is supposed to be. Japanese Wikipedia also has the lowest number of administrators per active user of all language editions. I think currently there are about 40 administrators, and English Wikipedia have about 1100. So administrators have the ability to block users, protect pages, edit protect pages, delete pages, rename pages, and so on. So they have a lot of power. And because Japanese Wikipedia is run by just a few dozen administrators, they pretty much control what goes on in the platform. The Japanese Wikipedia has never worked as intended because Wikipedia was created with Western values without regarding the differences in languages and cultures. And I can summarize my investigation in one sentence, and that is this. Wikipedia works only if the crowd is large, diverse, and independent enough to reliably weed out assertions that run counter to fact. And I think English Wikipedia largely meets this criteria. The crowd is large, editors community is large, diverse, and mostly do well with reading out misinformation and disinformation, but as it turns out many non-English versions of Wikipedia do not meet this criteria. So they don't really work. And after the article was published in Slate, I spoke to the Wikimedia Foundation representatives about this problem that I wrote about. And here's what they said. They said that they weren't aware of the problem that I had raised in my article, even though the problem that I'm talking about here seems to be going on for over a decade. So it's not a new problem, but they said they weren't aware of the problem at all. They have had a little contact with the Japanese Wikipedia community over the years, and there is no one on staff who speaks Japanese. They have about 500 employee staff at the Foundation, but no Japanese person, so they don't know anything about the Japanese Wikipedia. And the last part is something that several people from the Foundation repeatedly said to me. It's difficult to discover issues like this with more than 300 different languages on Wikipedia. So, you know, it's just really hard to monitor these languages, and we are a small organization and so forth, and my question is, so then why do you do in so many language versions? Which led me to write another article focusing on the how to fix part. Okay. And this article was published in Undark Magazine. I proposed creating a global Wikipedia. If the Foundation cannot ensure the quality of all of its language versions, perhaps they should just make one. One shared authoritative volumes of pages that users from all around the world can read and edit in the language of their choice. So, if you're Japanese, you can read and edit in Japanese using machine translation. If you're French, you can edit and read in French using machine translation. So, in other words, we are looking at the same page. When you look at the Nanji Massacre page, whether you're Japanese or French, you're using machine translation, but you're looking at the same page. This will be a technological feat, but probably not an impossible one. Machine translation has been improving so quickly in the last few years, and it has become part of everyday life in many non-English-speaking countries. If you're an English speaker, you probably don't use machine translation every day, but I think that, for instance, Japanese people use it almost daily. And I would have to say that I used Google Translate, and in the last few years it improved so much that it's pretty good. So, I don't think it's impossible to imagine something like this. But the reason why I suggested this, instead of like small changes here and there, is because the problem is built into the system. What I mean by that is, even if Wikimedia, Wikipedia may have managed, belatedly, to remove abusive editors from its Japanese or Croatian language versions, the same thing can happen again. In fact, I think the same thing will happen again, because the problem is built into the system. So, in order to really solve this problem, it requires a change of mindset. In other words, we need to change the way we think about Wikipedia itself to solve the problem. Lastly, I want to end this presentation, which is kind of early, but I was hoping to have some discussion with you if you have any questions and so forth. So, this is my last slide. Importance of diversity. The staff and the Board of Trustees at the Foundation are mostly Americans and Europeans. If what I described to you today was happening to the English Wikipedia, I am sure they would have noticed it and tried to address it a long time ago. But the Foundation has no one who speaks Japanese or understands Japanese culture, so the problem has gone unchecked for many years. And I think that this is why the diversity is really important. In tech, I'm not really working in the tech, but I have been to many conferences, and I hear that diversity is a big topic in tech, right? But we hear about diversity in terms of race, gender, sexual orientation, maybe age, and so on, but we rarely talk about diversity in terms of cultural backgrounds. This seems strange to me because technology platform like Wikipedia is used all around the world, so people in different countries and different backgrounds should be represented in companies and organizations. This is all solutions created by people of a similar language and background can miss important implication for how the product gets used by other people. And I think that the problem that I described to you today is an example of that. So I went through it very quickly. Is there any question or your thoughts on this matter? Yes. Well, one of the issues that actually Wikimedia Foundation told me about this kind of approach, actually they are actually doing that already with Wikidata, Wikifunction, which is using data, machine-readable data to create language in neutral one version. So they are kind of doing that already, but the reason why they don't want to do actually this global Wikipedia is because if they do that, the Western version will become dominant because most people are U.S. and Europeans. So U.S. and European version of history or other things may become dominant, and I think it's a valid concern. However, I would argue that from my perspective, it's already colonial. It feels colonial to me because the system itself is based on Western values. There's something that I described earlier. It's based on Enlightenment values. Again, Enlightenment is a great values that I personally agree with, but our culture is so different from that, and this is why the system doesn't really work well. So I do feel like it's already a Western system. Yes, I think what you raised is a valid concern, but I suggested this, not necessarily because I believe in global Wikipedia actually, but I think it raises the right questions. Thank you for this. Is there any kind of maybe middle ground of translating? I think there's such an initiative that they're trying to do, where it makes it easier for Japanese editors to translate English material. Some people have tried that, but then again, there's a group of a gatekeeper that is monitoring certain pages so they cannot really change that. Wikimedia Foundation said that they're going to look into it, this problem that I described, but nothing really has happened. Any other questions or comments? Anybody has ever edited Wikipedia? Oh, you have. How was your experience? I think that a lot of English Wikipedia editors have told me that they have actually seen a similar pattern of bullying and harassment that sometimes happens in the community as well, but not to the extent of some of the non-English Wikipedia where a certain group has been around for many, many years. They have learned the way to get around rules and abuse editors, abuse other editors. I came here to present today because it seems that Wikimedia Foundation really responds to some social pressure. Their approach is very hands-off. We do not interfere the content, we do not interfere with the community, but I do think that they need to do more. If there is more social pressure to do more, I think they will respond to that because it's a non-profit organization. Thank you so much for coming today.