Instant Genius - Wikipedia founder Jimmy Wales on what AI and Elon Musk mean for the site's future

Episode Date: October 27, 2025

Wikipedia. It needs almost no introduction. Few websites are more well-known, more commonly used and more foundational to the web as we know it than the online encyclopedia. This week, we’re joined by its founder, Jimmy Wales, to talk about how the platform has evolved over the past two decades, the challenges of maintaining trust and neutrality in an age of misinformation, and how AI could shape Wikipedia’s future. Jimmy also shares insights from his new book, The Seven Rules of Trust, exploring what it really takes to build credibility – and why it’s more important now than ever before. Learn more about your ad choices. Visit podcastchoices.com/adchoices

It's peak pollination season, and my business is scaling fast. To keep the nectar flowing, I need a phone plan with top priority data speed. That's why I chose Google Fi Wireless. My connections stay strong even when the hive is buzzing. Plus, unlimited plans start at $35 a month. Now that's a deal that doesn't stay. Explore Google Fi Wireless plans today. Plus taxes and government fees.

Google Fi Wireless is not subject to data traffic deprioritization during times of high network usage. Study and play. Come together on a Windows 11 PC. And for a limited time, college students get the best of both worlds. Get the Unreal College deal, everything you need to study and play with select Windows 11 PCs.

Eligible students get a year of Microsoft 365 premium and a year of Xbox GamePass Ultimate with a custom color Xbox wireless controller. Learn more at Windows.com slash student offer. While supplies last, ends June 30th, turns at AKA.m.m.S. College PC. You said this place was steps from the water.

We just haven't found the steps yet. How much did we save? Enough. Enough to get lost. Or you could book a stay with Hilton. Welcome to your ocean front room. Just steps from the water. The Hilton sale is on now.

Book on Hilton.com or the Hilton app and save up to 20% to get the stay you expected. When you want savings, not surprises. It matters where you stay. Hilton, for the stay. This podcast is sponsored by name, audio and focal. Streaming has made music more accessible than ever, but true listening is about more than ease.

It's about quality. British audio experts name audio, alongside French acoustic specialist focal, combine handcrafted tradition with cutting-edge innovation and high-end materials, delivering digital precision with analogue warmth, so you can experience exceptional sound at home. Music just as the artist intended. Visit name audio.com to learn more.

Hello and welcome to Instant Genius, the bite-sized masterclass in podcast form. Each week you'll hear from world-leading scientists and experts talking about the most fascinating ideas in science and technology today. I'm Tom Haworth, trends editor at BBC Science Focus. Wikipedia. It needs almost no introduction. Few websites are more well-known, more commonly used, and more foundational to the web as we know it than the online encyclopedia.

This week, we're joined by its founder, Jimmy Wales, who talks about how the platform has evolved over the past two decades, the challenges of maintaining trust and neutrality in an age of misinformation, and how AI could shape Wikipedia's future. Jimmy also shares insights from his new book, The Seven Rules of Trust, exploring what it really takes to build credibility, both online and in the real world, and why it's more important today than it perhaps ever has been before.

So Jimmy Wells, welcome to Instant Genius. Great. Thanks for having me on. So it's a bit of an exciting time for you at the moment. You've got your new book coming out, which we'll get to in a minute. But we're also kind of fast approaching Wikipedia's 25th birthday, right? That's in January. That's right, yeah. So today Wikipedia is one of, if not the most well-known, well-respected websites on the internet.

Take me back to the beginning. How did this all start? Yeah, well, I was watching the growth of free software, open source software, as most people call it. And I seeing programmers coming together to collaborate in new ways, and I realized that this kind of collaboration online could extend beyond programming to all kinds of cultural works. Had the idea then for the encyclopedia and got started. And there was obviously a lot to learn along the way. But that was really the fundamentals. It's just seeing collaboration online. Because you started with Newpedia, right, and then it kind of morphed into, or you started kind of on the side, Wikipedia. What was the main differences there that made Newpedia not a success

and Wikipedia the phenomenon that it is? Yeah. So with Newpedia, I didn't know about the concept of Wiki, Wiki meaning a website, anyone can edit and a very open platform. And so we organized things in a very top-down way. You know, there was a seven-stage review process to get anything published, and we were recruiting academics. People had to send in their CVs to show they were qualified to write on a certain topic. And ultimately, it just wasn't very fun for people. And so people were enthusiastic about the vision, a free encyclopedia for everyone, but the actual getting the job done was too tedious and not something people really were passionate about. And then when we switched to the wiki system, and we did switch, as you say, it was sort of a side

project. We weren't sure if it would work or not. We just thought, well, let's give this little tool a try. We got more work done in two weeks than we had in almost two years. And it really unleashed that early community to be able to just start writing and checking each other's works and becoming Wikipedians. So since Wikipedia then became the main project in the kind of 25 years since, what do you think are some of the biggest changes that it's undergone, other than obviously it's phenomenal growth in size? Yeah, I mean, what's interesting is how Wikipedia has been stable for very long time. And, you know, in the early days, there were certain things that we always knew weren't going to be able to last forever. So the earliest software was a

great little pearl script called UseMod Wiki, which didn't even have real passwords. So you could log in and create an account, but you couldn't put a password on the account so anybody could pretend to be anybody. That was clearly never going to work, but it got us started. And then in the very early days, there was no blocking mechanism. So you couldn't block people from trolling or whatever, except for I could ban their IP address at the server level. Again, that was never going to scale. And so we had to develop the tools that the community needs to be able to control the space and make sure that everything is in the right way.

And then, of course, all of the ideas of Wikipedia, you know, they are quite simple, but they had to be created from scratch, you know. So simple ideas like, you know, reliable sources, like we want to use sources that are good quality. Well, okay, that's kind of obvious. So we didn't have to write it down before we got started, but eventually you have to go, yeah, by the way, actually, we do only use reliable sources. You can't just sort of pick any random web page and call it the equal of a published scientific article and things like that. So a lot evolved over time, but it wasn't in the sense of fundamental change. It was really in the sense of formalizing some things that hopefully are pretty obvious.

You mentioned one of the reasons why Newpedia didn't work out as a project was it wasn't really fun for people to contribute to. Do you think that that is what drives Wikipedians to be a part of this community? You know, what is it that makes it so popular? Why have we got millions of articles now? So I always think it's two things. And I do think fun is a part of it. Just the idea that you're going to edit Wikipedia, you're going to meet some really nice and interesting people. Typically, people who are interested in the same sorts of things.

you are. So if you're super geeky about, I don't know, the history of cricket in the 1940s, I'm just looking here in this room I'm in, there's the wisdom, whole collection of wisdom, cricket, whatever it is, Almanac. If that's what you're obsessed with, then you're going to start editing about that, you're going to meet other people who are interested in the same thing. So that fun is part of it. But also just the charitable motive or the idea of like, actually, this is useful to the world, that you could spend an afternoon editing Wikipedia or you could spend your afternoon, I don't know, playing Grand Theft Auto. Probably at the end of the day when you go to bed at night, you're going to think, oh, yeah, world's a little bit better. I did something useful

today, updated four or five Wikipedia entries, as opposed to thinking, yeah, well, that's a, that's, that was fun playing the game for six hours, but I haven't really improved the world at all. So, I think it's both of those things. And obviously some people, it then becomes quite a huge part of their life. I'm thinking of people like Stephen Pruitt, I think, is the Wikipedia who's contributed sort of over six million edits. Have you ever met him or interacted with him at all? I have met him.

I don't know him well. He's very prolific editor. Of course, his edits style is lots of little edits, so lots of things like that. And we try to discourage what we call edit count. Titus, looking at how many edits you have as a measure of, you know, who's the most active. Obviously, with 6 million edits, you're very active Wikipedia in, but there are other people who edit much more slowly. You know, they go off and they do research for a day and a half and then they write a paragraph,

and that's also really important as opposed to just maintenance and, you know, using scripts and things to keep the site. So all types are needed for, you know, putting together this compendium of knowledge. Yeah, because I did, but I think he's contributed to something like, you know, over a third of the articles on there. But I guess that's not a problem then, is it? If he's doing these kind of small ones or is it? No, I don't think it's a problem at all. And yeah, I mean, sometimes I see, you know, this weird sort of like, and one third of it was written by one guy.

Well, no, one guy's touched a third of it, which is amazing, absolutely incredible. But, you know, he's mainly doing a lot of grammar fixes, spell checking, that sort of thing, fixing up references. you know, the technical details of making it nice, which is wonderful. So your new book, The Seven Rules of Trust, in your own words, kind of, what is this book about and why do we need this book now, do you think? Yeah, so, you know, we've seen in the last 25 years, and Wikipedia's coming upon 25th birthday soon, a real decline in trust in society, a decline in trust in media, in politics, in each other to some extent.

And meanwhile, Wikipedia has gone from being, you know, something people consider it as kind of a joke to one of the few things that people trust. And at the same time, Wikipedia is very trusting. You know, you can come to Wikipedia and get started today. You just click at it and start editing, and it's very easy to participate. And that's surprising, given all the toxicity that we see online

from social media and things like that, places that aren't really kind and thoughtful and respectful and trying to create something of value. So it felt important to say, okay, you know, we're seeing this increasingly lead to bad consequences in reality where, you know, politics has become violent in many cases. And the ability, you know, right now we see the U.S. government is in lockdown, shutdown because they can't agree a budget. And the idea of bipartisan cooperation is become very thin to non-existent in some countries. And so it feels important to say, hold on a second. Like actually, human beings are capable of a better way of living,

of creating environments that are trustworthy, that are trusting, where we can have genuine dialogue about ideas and be thoughtful and kind to each other. So that's why it felt important to talk about what are some of the things I've learned. Like, how does this amazing thing like Wikipedia, even work? How can that possibly work in a toxic world like this? Because hopefully people can come away with some useful messages that, you know, might be helpful in your personal life or in business or whatever it might be. Yeah, because, you know, in the book, so you mentioned social media. I think you describe it as across tribal lines, a place where trust goes to die. And then, obviously,

you kind of refer to Wikipedia in many ways is a bit of a ray of light with trust on the internet. do you kind of see these two sides of the internet as in competition with each other? Do you think that Wikipedia can compete with that kind of untrustworthy world that seems to be boiling over? I mean, I think we can definitely compete with that. I'm not sure. I think of it so much as competition, not like a business competition or traffic competition or anything like that.

Clearly, there is a need broadly for a shift in culture away from. culture wars to a more thoughtful, reflective way of dealing with each other and accepting people with differences and having more dialogue across different lines. I think we will always have both. I think it's quite normal that in human society, you know, I actually, I liken it, whenever I hear people talking about young people today and like, oh, they have no attention span, they're just on their TikToks and whatever. And I'm like, yeah, clearly, Clearly, there's this addictive short-form content that the young people do love. But guess what?

They're also binging 14 straight hours of very complicated TV shows. And so both are part of human experience and both sort of fluffy social media that's not actually, you know, particularly healthy for you. And serious research and thinking online, both of those can coexist. It doesn't have to be one or the other. but hopefully, you know, don't only eat junk food, have some nutrition as well. It must be weird, though, to kind of find yourself in the position of somebody who created this thing, which essentially is a compendium of facts and human knowledge,

when at the same time kind of from the upper echelons of political leadership, particularly in the US, some people might say, is kind of a lack of truth and trust is emanating. Are you kind of worried that despite the good work that Wikipedia might be doing that it's kind of being lost in the noise? I mean, to some extent, you know, I think we're actually living through a phase. And, you know, if I'm wrong and we're descending into a new dark ages, that's not going to be a great outcome. But I think that the political climate being particularly toxic is something that's not sustainable. Because, you know, you can't make good decisions in that kind of a climate. can't move society forward in any useful way.

At the same time, you know, it's always, it's tempting but a mistake to oversimplify because there's also an enormous amount of incredibly good trends in society and things that are actually quite good that are happening. So, you know, it's a bit of a mess at the moment, but I'm optimistic. So your fifth rule in the book is kind of about civility and treating others as we would want to be treated or how we would teach a toddler to treat others. us leading on from what we've just been talking about, how is it that you think Wikipedia has been able to maintain that kind of relatively constructive culture when the rest of the internet

is kind of a bit of a dumpster fire? Yeah. I like that you say relatively constructive, which does acknowledge, like, obviously we have our quarrels internally and, you know, people do get upset sometimes. That's just part of the human condition. I mean, I actually think it's part of the design. So it's partly software design. it's partly what I call community design.

It's really a combination of things. So, you know, if your business model is based solely on ad revenue and clicks and engagement, then it's really easy to actually, I would say, accidentally go down a very dark path where you're promoting content that's not making people happy, but just keeping them clicking for a little while longer and a bit angry and arguing with each and, you know, that's very unhealthy in the long run. Whereas Wikipedia, like, we're, we don't have any ads and we're supported by donations from the general public.

And so, you know, our business model, if you want to call it that, I mean, we're a charity, but we do, you know, have a business model, I suppose, is, you know, we don't, we don't need you to be on the site addicted and clicking as much as possible. We need you to remember us once a year when you think about donating money. And when we send you an email saying, oh, you don't. the last year, have a, you know, have a look. Maybe you could donate again. That would be great. And you will if you think, yeah, Wikipedia is good. It's like, it's helped my life. It's been meaningful to me. It's useful. It's trustworthy. Even if it's not perfect, it can still be

trustworthy. You know, so that drives us in a different direction when we think about the design of the site, when we think about the rules of the community. You know, the importance of civilized discourse is like super, super important because if we don't have that, then we Wikipedia ends up being full of nonsense. So, yeah, I think it's really a lot of different fundamental pieces of the puzzle. You mentioned that, you know, Wikipedia is generally trustworthy. Obviously, it's not perfect. I think most people, you know, trust Wikipedia to a degree, but probably not enough to kind of cite it in an essay. I know when I was at school, for example, that you weren't allowed to do that. Do you think that that's the right way to use it as a resource?

do you think that that's the kind of level of skepticism you should perhaps have to it? Yeah, I definitely absolutely think that's the right way. And of course, I mean, I always remind people, like when I was young and in school, we weren't allowed to just quote encyclopedia Britannica, not that there was anything wrong with Britannica, but that isn't what an encyclopedia is for in the research process. It's to give you a background and get you oriented. And then in our case, you follow those links and look at the sources.

That's what you really should be doing. I mean, I would make an exception if you're, you know, you're 14 years old and you've written something and you've bothered to put in a footnote, hey, that's progress. You know, you didn't just rip it off. But, you know, once you're an adult or, you know, at the university level, you really, you know, that's not, being a sightable source isn't really the goal. It's a goal of helping people be productive and get oriented and get competent to write something or to do some research and things like that. So yeah, I think that's fine. When you need to build up your team to handle the growing chaos at work, use Indeed

sponsored jobs. It gives your job post the boost it needs to be seen and helps reach people with the right skills, certifications, and more. Spend less time searching and more time actually interviewing candidates who check all your boxes. Listeners of this shell will get a $75 sponsored job credit at Indeed.com slash podcast. That's Indeed.com slash podcast. Terms and conditions apply. Need a hiring hero? This is a job for Indeed, jobs. This episode is brought to you by Subaru. Go further in a long-range Subaru hybrid with up to 581 miles per tank in the Forrester hybrid. Longer range, better fuel efficiency, and legendary symmetrical all-wheel drive, standard. The Subaru forester hybrid. Visit Subaru.com slash hybrid to learn

more. Maximum range based on EPA estimated combined fuel economy and a full tank of fuel. Actual mileage and range may vary. This podcast is sponsored by name, audio. and focal. With over 100 years of combined expertise, Name and Focal have been bringing music to listeners just as the artist intended. Since day one, this mantra has shaped every innovation in high-fi design, technology and acoustic engineering, balancing craftsmanship and tradition with pioneering thinking. Name Audio pushes cutting-edge technology to ensure digital precision whilst sustaining Pratt, pace, rhythm and timing. the elusive quality that makes music feel alive and gives it emotional texture. Today, in partnership with French acoustic specialist focal,

name audio creates systems that deliver exceptional sound and unforgettable listening experiences at home. Try it for yourself at a focal powered by name boutique. Visit focal powered by name.com for more information. So your rule number seven is all about kind of how transparency builds trust. In that spirit, what do you think over the years have been Wikipedia's kind of biggest mistakes? Yeah, I mean, it's one of the reasons I struggle with this kind of question is because I'm always a pathological optimist, so I think everything is fine always. But that's clearly not.

I say pathological, because I know it clearly isn't true. You know, we could go into specific mistakes, you know, articles that were wrong or things like that. But that's just part of the way things flow or time. and you can't really regret that too much. There are other cases where, you know, I can say we, you know, at the Wikimedia Foundation, made changes to the software without really checking with the community first and they went crazy and they were right in the end. Those are the kinds of things.

But broadly, you know, it's been a good ride, you know. There's always something new, something to learn. You know, there's always articles that need improving and sort of areas and big debates and things that we need. need to have. But I think if we ever lost that spirit of saying, well, yeah, Wikipedia, yeah, it's pretty good, but it's not good enough yet, not from what we really want. Let's dig in, let's see. And when we get criticism, and I think this is really important, it's important to not just be defensive and stick your head in the sand. It's actually important to go, well, hold on a minute, let's check. Like, let's take this seriously. Sometimes you'll look and you'll go,

yeah, actually, you know, when some crackpot says something's crazy in Wikipedia and you look and you're like, yeah, crackpot, not that interesting. And in other cases, you're looking, you're like, yeah, actually, this is pretty one-sided and it actually should be improved. And I think that's really important for us to do. Yeah, in the book, for those who read it, you are pretty candid about some of the kind of specific cases and mistakes that have been made over the years. One thing you kind of mention is the fact that, you know, any individual or politician or organizations can kind of go through and scrub rule to facts. And that can be quite difficult to. But, you know, you know, pick up at times. Has Wikipedia solved that as a problem or can it never really be solved?

I mean, I think it's never been that big of a problem and it's always something we have to think about and deal with. So there's a lot of, you know, work goes into checking the sources, validating things. It's pretty common that someone will come in and start trying to make a biography one-sided or something like that. And then the community has to step in and go, no, no, no, that's not how we do it, and you can't just come in and wholesale delete criticisms and things like that. And obviously, that's an ongoing thing, but there's no magic solution to that if you also want to remain open to criticism. Like if somebody is saying, hold on, this article is not very good, I'm going to make some edits and change it. Well, sometimes they'll be right and sometimes

they'll be wrong, and we have to stay open to having that discourse and that dialogue. So, you know, by and large, it's always been a part of what we do. And we try to educate. people about, you know, what's the best approach? If you've got a problem, you know, let's hash it out on the talk page, not just, you know, edit war in the article. That's not really helping anything. Wikipedia, perhaps now more than ever, is facing criticisms from certain places, you know, Elon Musk calls it wokepedia. What's your initial reaction to kind of those criticisms that particularly seem to be coming from the political right at the moment, I suppose you'd say? Yeah, I mean, it isn't great. I mean, you know,

when Elon calls us Wocopedia, he's just wrong, just factually wrong. Like, that doesn't make any sense whatsoever. I remember, you know, there's this sort of a gotcha question on trans issues. Like, what is a woman? And the, if you go to Wikipedia, the entry on woman, you know what it says? Adult human female, or adult female, human, whatever it is. It's completely uncontroversial, right? And, you know, people who want a, you know, people who can't quite come to that very simple definition, guess what?

It also goes on to talk further down in the article about some of the complicated questions of gender and society and so on, all very valid. But the idea that we're, you know, have become sort of crazy, less wing activist is just incorrect, factually incorrect. That doesn't mean there aren't areas where we should improve. There are, you know, there's definitely cases where I look at an article, I'm like, oh, this doesn't really, it's not fair to both sides. And we need to think about that and we need to work on that. But, you know, the right answer is to say, yeah, get more people involved. You know, I would like more kind and thoughtful people who see a bias in a Wikipedia entry to not assume that it's caused by some crazy woke activists who are going to block you the minute you disagree. But to say, actually, people are just writing from sources and they may not have seen the other. side of the debate and we can bring that in and chew on it together. Yeah, and you kind of highlight the important point there because it could create more of a split and then you would end up getting bias if only sort of left-wing people are the people who trust Wikipedia and only right-wing people go elsewhere for that facts and information. Yeah, no, I think that's one of the, you know,

it's like if somebody is a very kind and thoughtful, conservative, I would love for them and intellectual, love them, come and join Wikipedia, come and help us out. And if somebody is a crazy woke left-wing activist who's here to crusade, I'm like, oh, you're going to be so tedious and annoying. Like, please don't, you know, don't think that because Elon calls us Wocopedia that you found your new woke home, like, that's not how it works. Like, we're not trying to be an activist site in that sense. We're just trying to document the world and be fair to different perspectives and viewpoints, and that's really important. Do you think there is something inherent about sort of the Wikipedia community that maybe perhaps makes them lean slightly more one way than the

other. I guess, like, for example, you'd be remiss to say that university culture isn't more left wing than right wing. And so perhaps there is a kernel of truth in people on the rights arguments about what's going on there. Is there something similar like that going on at Wikipedia, do you think? I think, yes, of course there is. You know, when you look at any sort of body of work, so not just university, but also, like, one of the things to think, about is Wikipedia is global, English Wikipedia is global, and, you know, people who speak English across Europe would, I would say, if you did some sort of measurement on various political issues and so forth, would tend to be more to the left than the average person in

the U.S., which means that if you're on the right in the U.S., you may see things that surprise you or that you hadn't considered because you haven't really had exposure to broader, wider culture. That doesn't make us biased. It makes us more inclusive in sort of describing more arguments and hopefully doing so in a fair way. But of course, people will perceive that in a way that they will perceive that. And that's valid. One of the areas of the trust landscape, I guess you could call it, that's changing most rapidly at the moment, is all to do with AI, large language models. A lot of these models are trained on Wikipedia. What's your kind of initial thoughts on Wikipedia being used for that? Well, I mean, the first thing I always say is,

thank goodness that we don't have AI models that are basically trained on Twitter alone. That would be a really weird and angry model. Hopefully having training materials that are fact-based, thoughtful, reflective, that's actually really important. And we're, you know, the point of Wikipedia, it's completely freely licensed and people are allowed to do whatever they want with it. They need to follow the conditions of the license, which is attribution, and that's something we do talk about. But broadly, you know, I think it's a good thing if we have AI large language models, which are more fact-based. And in fact, one of the big trust issues that AI has right now is the hallucination problem is unsolved.

And some would argue, Gary Marcus is a well-known AI researcher who argues that it's unsolvable with the current technology, that the very nature of the way large language models work implies that they will hallucinate. And he thinks that the solution is going to lie somewhere else. And he's sometimes portrayed as a detractor of AI. He's actually quite an AI enthusiast. But he thinks large language models are maybe a dead end. So that's all very interesting. I mean, I'm not an expert in these things. I'm an amateur expert, I guess, because I'm obsessed with it and read about it all the time,

but I'm not an AI scientist by any means. For me, what's really exciting about AI is the potential that we might find some ways to use AI to support our community in Wikipedia. So what I mean by that, you know, things like I wrote a little thing to, you know, I can feed in a Wikipedia entry and a short entry because the context matter, so you have to keep it kind of shorter. The model loses its train of fault. But short Wikipedia entry and five sources, let's say, feed in the sources, feed in

and say, is there anything in the sources that could be in Wikipedia but isn't? Or is there anything in Wikipedia that's not supported by the sources? And it turns out it's pretty good at that. Like it will say things like, oh, well, three of the five sources said the person, was born in Munich and Wikipedia doesn't mention that they were born in Munich. Probably you could add that. You're like, oh yeah, probably I could add that. That's pretty good. That's very primitive. That's just me playing around. But we have a machine learning team at the Wikimedia Foundation who are really investigating, you know, what are the ways that we can use open source models to help the

community do things that they find useful? This summer, I talked to a French Wikipedian at our annual conference. We do a conference every summer. We were in Nairobi this year. And she goes through French Wikipedia looking for dead links, because links eventually die on the internet as websites change or whatever, and she finds new links. And I said, oh, and it's, you know, it takes a long time. It's a lot of work. That wouldn't it be interesting if, you know, you don't need an AI to find a dead link, you just find a dead link, but then what if the AI could go out and go to some typical sites where we would find sources, grab some sources, quote the sources, quote they'll say, oh, this source seems to support that fact where the Deadlink was supporting that fact,

is this good enough? And she's like, oh, wow, that would be very helpful, right? And that's something that was unthinkable five years ago. Like the technology didn't exist. They could do some keyword matching. But to really have it say, yeah, this seems to support that fact, that's new and interesting. And if that kind of work, which is just maintenance work for Wikipedia, which people do a lot of, It's sort of like, you know, she might compare it to, I'm not sure she would, but some people have compared it to knitting. You know, it's sort of something you do with a little, you know, past time and it's fun and it, you feel good after you've made something. So I think people might use tools like that, and those could be very exciting in terms of helping us elevate the quality. And that isn't about AI taking over the job of writing Wikipedia.

That seems like a very far-fetched concept right now anyway. You don't think sort of with the pace of change that that could be something. you know, in the next five years or so that could be happening? Hard to say. I mean, it's possible. I mean, it's not an era where you can say, well, I will never be able to tell a really bad dad joke, but boy, it's actually very good at telling really bad dad jokes. So, you know, like I wouldn't want to say absolutely never,

but at least not in the short run. I mean, the latest models are still from a Wikipedia perspective, just nowhere near good enough. humans still holding out, that's good to hear. Outside of Wikipedia, in your kind of classically optimistic style, you also think that AI could have some positive impacts for trust more broadly. I wonder if you could dive into some of the reasons why you think that could be. Yeah, so even though you said outside of Wikipedia, I'll give a Wikipedia example,

which is in one sense, we're very, very transparent because you know, you want to understand the rules of Wikipedia or you want to understand how those rules are made. it's all public. It's all, there's, you know, and you go to any popular page of Wikipedia and go to the discussion page and you'll find archives of dozens, hundreds sometimes of past discussions, so you can understand every decision that was made. Well, that's super transparent in one sense, but it's also not transparent in the sense that who's got time to read 100 pages of back argument and story?

And so maybe AI could help summarize some of that stuff. And if it can do it accurately enough, so you can say, gee, why does this article say that? What are the debates that were had? It could probably do a decent job of that, and that might be very useful. And obviously other organizations, companies, you know, who have a desire to be transparent, but find it very hard because it's quite expensive to publish reams and reams of data and information. Maybe they can find better ways to summarize. You know, actually, one of my big beefs, I'm a programmer, but not a very good programmer, and I use AWS. And one of the my beasts of AWS is there's so much documentation and it's actually

sometimes hard to find what you need and make sure you're not reading outdated documentation. And they now have an AI tool to help you with that. And it's actually kind of helpful. And so I think that that as an inspiration could be quite good, the ability to summarize things in a useful way. I mean, a lot of people are using, I don't use it because I rely on other people to be my memory and take notes and things. But, you know, AI summaries of meetings apparently have become very popular because, you know, you can sort of have that AI summary. And a week later, you can go back and go, yeah, what do we talk about in the meeting? Oh, yeah, yeah, the AI summary is decent. That kind of stuff sounds useful to me. And I think that's, you know, anything we can do to

make our lives easier is generally always going to be a good thing. I'm curious about, because there are some unintentional biases on Wikipedia, things like. like there are, you know, articles about men tend to be longer and they tend to be more career focused rather than women tend to focus more in relationships. Do you worry that sort of, you know, with AI, what you feed in is what you get out? So if LLMs are relying quite heavily on Wikipedia as a resource as probably the biggest bank of human knowledge on the web there is, that we could be enhancing those biases into the future. I do.

I think that is something that really needs to. be focused on, particularly the people who are training AI models. I need to give that a lot of thought. I mean, the two largest sources of data are, for most LLMs, that stats I've seen is Wikipedia and Reddit. And both are, you know, they're great and they're flawed and they are what they are, but they're also both very heavily male-dominated areas. And hopefully with Wikipedia, because we have this commitment to neutrality, we're already

sort of trying to be conscious about that sort of thing. But if you aren't careful, you would just sort of blindly go into reinforcing existing biases. And at the same time, though, because the large language models can be trained and they can be prompted and they are, you know, they have the capacity to do that sort of thing, it isn't hard to imagine having an AI that runs constantly over Wikipedia that, you know, you ask it to say, look for instances of thus and such type of bias and call to our attention an entry or, you know, something that we could look at. And as long as it could be used to help the humans to say, oh, wow, you know, nobody really realized, you know, that articles about Nobel Prize winning

economists tend to mention the families of the women who've won and not mention the families of the men. okay, that's interesting, right? And it may be something that nobody noticed. And maybe the reason is because the media gets very interested in, I don't know, let's do a UK example in Theresa May's shoes. That was a big news thing. And it's like, and nobody cares about Rishi Sunak's shoes. Maybe they did, I don't know.

I'm not sure if they did. I don't know. He's really rich. He probably has a really nice shoes. But anyway, nobody paid attention, you know. So like maybe it's not entirely Wikipedia's fault, but like if an AI can help us to notice it, then we can go back and go, okay, hold on a second. Let's think about this a little more, you know, more carefully and let's dig a little deeper on how we think about these things.

At least there's a possibility. But yeah, anything, I think this is broadly true, my approach to life, anything you can do with kindness and thoughtfulness and reflection is going to be better than anything you do kind of blindly and impulsively. And so that's sort of the broad Wikipedia answer to Twitter. I was like, be a little better off, but we just all slowed down and thought for a while. So that was Jimmy Wales, founder of Wikipedia. If you'd like to learn more about the topics we discussed today, then why not check out his latest book, The Seven Rules of Trust, which goes on sale on the 28th of October. Thanks for listening to another episode of Instant Genius, brought to you by the team behind BBC Science Focus. Please consider subscribing

to Instant Genius on your preferred podcast platform. Or, if you'd like to watch to our presenters and guest speaking in person, then you can also check out our YouTube channel at ScienceFocus. The current issue of BBC Science Focus magazine is out now. Pick up a copy wherever you buy your favourite magazines or downloaders on the app store of your choice. You can also find us on Apple News or online at sciencefocus.com. This podcast is sponsored by Name, Audio and Focal.

The texture and emotional depth of music can be lost through digital sources or poor signal. Name Audio believes you can have digital precision with analog warmth. Alongside French acoustic specialist vocal, Name creates high-end audio systems combining innovation with craftsmanship, so you can listen to music, just as the artist intended. Discover more at name audio.com. Ambition comes in all shapes and sizes.

At First Citizens Bank, we roll with your goals, because we're built for what you're building, fit for your ambition for citizens back. Enjoy more ways to save at Ralph's, like low prices in every aisle. And when you download the Ralph's app, you can clip and save more with digital coupons every week. Plus, you can earn fuel points to save up to $1 per gallon at the pump. At Ralph's, you can enjoy more ways to save and more rewards every time you shop. So it's always easy to save big every day with savings and rewards.

Ralph's SoCal for over 150 years. Savings may vary by state. Fuel restrictions apply. See site for details.