 Now, we have another session, which is the Wiki Women Open Discussion, Language, Diversity and Gender, to be led by Tila Capolaro, who is online, and Masana. Thank you so much, French, and we are incredibly efficient with some time to spare, so we have more time for discussion. Thank you so much to our panelists for making space for everyone to participate. Tila, for colleagues who are online, nice to see you. Thank you so much, Rosie. I was actually going to run so that you can be comfortable. Tila is in the chat room answering questions. I've been told that we will have a summary of questions pulled up on the screen, and whilst we're getting our summary of questions from online, we would love to hear from you, all of you. Everybody has to say something at least once today. Any thoughts and reflections on the presentations, questions for our panelists, for our wonderful organizing team, reflections on Wiki Women Summit? Yay! I did have a question about the speedy deletion process. How actually easy is it when you get a situation like that to get images speedy deleted from Wiki Commons? Obviously it's very important to happen if someone wants their image speedy deleted certainly in a situation like that. I'm just wondering whether the processes make it easy for you to do that. Does anyone have any? Yeah, Sanchin and Mariana, can we turn to you for that question, and can folks raise their hands? Rosie and I are just going to move through the room and make eye contact with people to get participation. How quick is the speedy deletion process in your experience? Well, not as quick as we would want it to be, but we have a resource on how to request or ask for speedy deletion. I think it also helps to know some people, and I think that we are happy to share what that protocol and that link to that process is, but we wish it was much quicker. Sometimes it could take hours, could take days depending on what you want to be deleted, but we do have a resource that we can share. Thank you. That would be fantastic, and I definitely confirm that we have had experiences of people not even knowing who to ask, and I think that's one of the ways in which we can build up safeguarding in our movement, which is how do we make it easy for women and non-binary people to find information when they need help, and how do we build up a body of volunteers and WMF staff who can step in to support speedy deletion. But it is a gap, and Marianna and Sunshine, if I could ask you to email that to me, and I can post that on the Twitter page. More questions, more comments. More inspiration. Yeah, thank you so much. I have a little consent regarding tools. I was reading a research about women. I think one of the community did a research about women, and they were talking about how women find the interface of Wikimedia projects unfriendly. It's like it's something that we don't like to do, but amazingly, there are tools that have been created. When you go to Toolforge, you see a lot of tools. But I was wondering, I was still looking for a tool that can help me to edit about women, because you could have tools that enables you to edit about particular topics, let's say history, maybe this country, monuments, like specific things. But I was looking for women, I wasn't finding it. So I'm wondering, I'm not a developer, I don't develop tools, but as a Wikimedia movement here, we're concerned about women. How are we trying to maybe work with this developer community to develop those kind of tools that can help us to edit about women? I can start answering the question, and then we'll have someone continue. The community called Women in Red has a resources center where there aren't what I would call tools, so I need someone to answer that, but there are essays regarding specifically things one might do when editing articles representative about women. So it's not only the biographies, but also information about women's works, broadly construed women's issues, broadly construed. It's not a tool, but it's information as to things you should be aware of, frequently asked questions, tips and tricks, things like that. Hello. I'm very impressed about the work that they presented from Wiki, Visible Women, by Whose Knowledge. We have this project called Gender and Cultural Diversity on Commerce and Feminism, and we are trying to understand the diversity of Wikimedia commons farce on commons, and we are currently working to build like a tool that will help us annotate random images to see how diversified images are. We already know about the gender gap on Wikipedia, but we don't know how it is on Wikimedia commons. So we are currently at the early stage of the research, and I'm really curious to know how you are working with tools like Ruby mentioned. Do you have any tool that you are currently working with in terms of your work? How do you identify random images to be able to understand what exactly they represent on Wikimedia commons as part of your work at Visible Women or Whose Knowledge? We are also calling on people who are interested in exploring or want to know about how we can help identify this diversity or cultural and gender diversity on Wikimedia commons. So we invited people to help us annotate random images and we're happy to learn more about what is working with other communities and how they've been able to navigate around these challenges to identify all these cultural diversity on Wikimedia commons. So I'm just using this as an opportunity to invite people who are interested in exploring with us, and also if you have any idea about how you can help us discover these findings, that would be very, very useful. Thank you. Thank you for the question. It's very interesting. We don't have a specific tool and so we need to talk and figure out together and plan it together. We know, we are exploring the tool called ESA, many of you know, I think that Wikilob's women have been using that tool for a long time now, and it's a tool for that you select a category on commons and then there is like a game, you see images and you can add descriptions, captions, Wikidata, distributed data items, and we really want to organize something like a data fest or a feminist data tunnel, something like that, for doing that exactly. So maybe we can speak and organize something together around that. But in terms of our analysis on diversity until now is not an analysis based on quantitative research. We have also observations. We document what we see and we try to speak about that, analyze that, but we really want to develop maybe including qualitative methods for analyzing the images, annotate with a feminist land, the images. So that's the work we are doing so far. We are far, but maybe we can do things together around that. We have some questions online. Yes, I will read them. Colleagues online, Sheila is there and copying and pasting your questions for us to be able to read them. So please feel free to engage in that space. The first one is from Aynali. In my opinion, the automatic image recognition should just be disabled as it often suggests two generic tags. It's not so much a question, but a reflection of the ongoing discussion on comments, talk, structured data. So thank you for that reflection. Any of our panelists or anyone else in the audience have a reflection on the generic uploads or related? Also curious regarding the first session that we heard Clifford and Colleen and Christine speak about regarding the use of maybe chat GBT to help with the work on Wikipedia. Does anyone else have that experience? And if so, how's that been for you? Is anyone trying to use chat GBT or what? Ah, we have one here. And can we just ask colleagues to stand and introduce themselves for our participants online if you're comfortable? Hi. My name is William and I'm from the United States. And on the topic of using chat GBT in work with Wikipedia, this is something that I have tested out on several occasions to try and figure out, like, how can we use this? And, like, what can this do mainly? Because the best way to figure out what something can do is to push against the limits of it and try to figure out how does it react to these situations. And I've often found that, like, a lot of people when thinking about chat GBT, as it is with Wikipedia, would simply think about, can we just use this to automatically generate articles? And, well, to that, the answer is yes, but with a lot of asterisks. Because it can either make stuff up or include it making up citations. So that's something that one has to think about. However, something that I've wanted to bring up here is the potential of using it for applications beyond just generating articles by itself. Like, for example, can you use it to find problems with articles or suggest improvements for articles? The first session in here earlier today kind of touched on that a little by discussing, can we use it to address biases, for example? And also, if you connect chat GBT with or other language models with the Internet, and can you use these to find sources for topics? So I guess the main thing to think about here is what are the possible applications beyond just generating articles by itself? And obviously, if we use these for these purposes, we can find a lot of benefits to the incorporation of AI tools. Thank you, William. Leta. Just starting on to what you said. For me, the main interest in using chat GBT was to find the knowledge gaps in existing Wikipedia articles, particularly in women's health. What I did was I copied the whole text of the article, put it up on chat GBT and asked chat GBT what was missing in that article. And it was sometimes quite good in finding out, saying that, look, it doesn't say anything about the clinical implications of this condition or it doesn't say about the anatomy of this particular organ. So there were some points which I could always take from chat GBT and try and think about it and incorporate into that article. But on the other hand, when I tried to make chat GBT write a full article about a particular anatomical structure or a particular disease, it was quite incomplete. It was always hallucinating and it just brought up a lot of, you know, made up citations from different things. I mean, there were so many things which did not exist in real life, but chat GBT was just using it. So I think it's useful when it comes to, you know, if you want to rephrase some sentences, you have got a particular text from a particular journal article which you want to rephrase into popular science or, you know, a language that people can, most people can understand. In that case, chat GBT is quite good at rephrasing your sentences. But to create knowledge, I think you always have to use your discretion and figure out if this is true, before you actually put that up on Wikipedia. So I think the future where chat GBT generates articles automatically, it's quite, quite far. I don't know if we'll ever reach there, but can chat GBT assist humans to make articles have more quality? Yes, I think that is possible. Hello? And I... Yeah. Come to think of it, I should submit a topic about chat GBT here. Okay, and I think it's... There are a couple of points that's quite interesting about chat, using chat GBT about this. One point is that I call it about... I call it design political correct. Design political correct. That means that, for example, I asked chat GBT about... and to discuss about the Disney movie Little Mermaid using a black woman as a leading actress. And its answer is so pleasing that outstand almost 80% of things that you saw on the Internet. Yeah, so that's why I call it... It has something... don't not be evil. So it has something... I say that design political correct. I think that is very quite interesting. It's quite good because if you are using it as a tool for students to learn something and for some... And that is one thing. And the other thing is that I try to use it to translate the article from English to... to translate Wikipedia article from English to Chinese. And along with... along with Google Translate, it works well. And I think it's good to be too... There are two copies of translation that you can compare them. But another benefit is that if you don't understand the translation you can still ask it about to explain it. Yeah. And you can still ask it. And there's a trick here is that because chat is trained in different language with different texts. So if you ask... ask the same question in English or in Chinese or in different in French it may get you different answers. Yeah. And that is a little trick you must consider. Yeah. Thank you so much everyone. Any... Yes. Hi. Please, my name is Ruby from Ghana and I'm part of the Open Foundation West Africa community. I'm very interested. I wanted to ask the follow-up question regarding what she's saying that if you ask chat's TTP in a different language it gives you different responses. Could it be that it picks references from that language category or... I just wanted to understand the disparities. Yeah. The questions we asked just TTP you said it changes when you ask different languages. And why is that? Yeah. Because you actually actually learn from a lot of public texts. Yeah. Try to really learn from... Actually actually I'm not quite... I'm not 100% correct because every time you ask TTP you'll give you a different answer. Okay. So... So even if you ask it in English twice you'll give you two different English texts. Last time I asked you about a question about transgender women and next time I want to copy it and ask it again, well, I cannot find the original answer. It's a generating machine. It generates an answer for you. So we don't have a standard answer there. But... The difference of different languages because you learn from a different language when it... that is caused by the course that you learn. In English you learn from the English text in this content all the free English text in the world. But in Chinese then you learn from all the different language free Chinese in the world. And there's a cultural difference between the free Chinese text and free English text. So you'll get free... you'll get different results. That's sometimes because and so sometimes I cannot get results in Chinese because most of the recent documents are... recent document... changing documents are in English. So I cannot get results in Chinese. And that is... I don't know if... does that answer your question? Great. We have about a minute left for any final reflections. Hi, I'm Natalia and I was sitting here and thinking that in 2011 I was in Gdańsk in Wikimedia and we had a Wikimoman lunch. One of the first one. And it was such a small group of us and we were kind of shy about it. Is it okay that we take this space, that we do it? And I remember the conversations were actually about how can we create a space with more diversity in this movement. I'm here at the Wikimoman summit and I'm a bit emotional because we kind of have this space already. We're not talking about can we have a diverse space in this movement? We have it. We celebrate it. And I'm listening to the sessions thinking that we are in such a different place because right now we are experimenting with how can we build diversity in many different ways through data, through images, through languages. We actually got here. It's like absolutely amazing. Just a last I wanted to recommend especially those random contests in like Wikimedia Commons contests. I would strongly recommend that we describe images in our languages like local languages. Especially when you are uploading images about women. Most of the time we just put the descriptions in English without putting the description in other languages. I think it helps a lot. One of the examples I found was Dagbani. I was looking for the name of something in Dagbani. I don't know the Dagbani name but I just typed it in English and then it brought the image because there was a description of that which has an image, the description in English and Dagbani. So I was able to see the picture and I could find the real name of that particular file in my language. So for somebody who like to work with language companies, I would strongly recommend that we try to include these descriptions in various languages, not just English. Thank you. Thank you very much. We're going to take a short coffee break now. There's coffee service for us in the main concourse where you got your name badges and come back after to see you at a quarter tail. One moment. Kelsey I thought Kelsey was saying something. Kelsey was making a motion. She says it's okay. Thank you. See you at a quarter tail.