Katherine Maher: Former CEO of the Wikimedia Foundation

Here's something remarkable and not in a good way. Disengaged employees cost US companies almost $1.9 trillion a year. That's a lot of zeros. Why? Because when teams lack the right skills, productivity declines, turnover increases, and strategies stagnate. But it doesn't have to be that way. At Project Management Institute, home of the PMP and globally recognized certifications, your people can learn to collaborate, deliver on time, and actually get things the cost of not upskilling way higher than the investment Project Management Institute. Find the answers@pmi.org.

I'm Guy Kawasaki and this is remarkable people. From 2015 to 2016, I was on the Board of Trustees of the Wikimedia Foundation. This organization oversees the 13 Wikimedia properties, including Wikipedia, Wiktionary, Wikibooks, Wikimedia Commons, and so on. Its goal is to provide the infrastructure that makes free knowledge available to everyone. There are more than 200,000 Wikimedia editors and contributors. More than a billion unique devices access Wikimedia projects every month. Wikipedia alone averages approximately 18 billion page views per month. The English version of Wikipedia contains approximately 6 million articles and people add approximately 600 articles per day. In 2020, supporters donated more than $120 million to the organization. With first hand certainty, I can tell you that leading Wikimedia is one of the hardest jobs in technology. This is because you have to make so many people happy. Wikipedia editors and contributors, end users and consumers, donors, partners and employees. And your tools do not include offering employees stock options, selling advertising, or collecting affiliate fees. This episode's remarkable guest is Catherine Marr. Until April 2021, she was a Chief Executive Officer and Executive Director of the Wikimedia Foundation. She did a remarkable job in this remarkable position. She has a bachelor's degree from New York University as well as extensive education in the Middle East. She worked for UNICEF, National Democratic Institute, World bank, and AccessNow. She was the Chief Communications Officer of Wikimedia from 2014 until 2016 when she became the Interim Executive Director. Something that truly irritates me is that teachers tell students not to use Wikipedia as an information source. So this interview starts with that issue. FYI, imho, Wikipedia is the best source of information in the world. This episode of Remarkable People is brought to you by Remarkable the Paper Tablet company. Yes, you got that right. Remarkable is sponsored by Remarkable. I have version two in my hot little hands and it's so good. A very impressive upgrade. Here's how I use 1. Taking notes while I'm interviewing a podcast guest. 2. Taking notes while being brief about speaking gigs. 3, Drafting the structure of keynote speeches. 4, Storing manuals for the gizmos that I buy. 5, Roughing out drawings for things like surfboards, surfboard sheds and office layouts. Six, Wrapping my head around complex ideas with diagrams and flowcharts. This is a remarkably well thought out product. It doesn't try to be all things to all people, but it takes notes better than anything I've used. Check out the recent reviews of the latest version. I'm Guy Kawasaki and this is remarkable, people. This interview took place during her last week at Wikimedia. I'm the second to the last interview. Is that true? Tomorrow's the Daily show. And that's it. So it's Guy Kawasaki and Trevor Noah in the same breath.

That's right. That's right.

I have arrived saving the best.

Yeah, yeah.

Was Oprah busy?

I think yes. I think she kind of did her one interview for the year. Yeah.

Well, we don't want to talk about how the royal family has treated Wikipedia. Let's just say you're at a family dinner and your cousin is a teacher, and she says to you, I never let my students use Wikipedia because anyone can say anything. You can't trust Wikipedia as a source. What do you say to her?

I'd say, oh, that makes complete sense that you'd have questions. And of course, your students are probably using Wikipedia anyway. And so I'd love to talk to you about how you should teach them how to use Wikipedia in a way that helps them learn.

And then what?

And then what? I'd say, look, all the kids are using Wikipedia. We all know it. I was never allowed to use cite an encyclopedia when I was writing essays, even when they were like hardbound versions in my school library. It's a great place to get started learning. It offers all those wonderful citations at the bottom for kids to do further research once they've gotten some basic context. And it's also an opportunity to teach young people about questions of digital literacy, media literacy, how to do research on the Internet, what information to trust, how to be a critical reader. And so instead of being afraid of Wikipedia, maybe the best place is to embrace it and use it as a teaching tool for the students who are, because they're going to use it anyway.

And then your cousin says, but, Katherine, anybody can change anything. How can they believe anything? And you say.

I say, Wikipedia is mostly accurate. Wikipedia, sort of, statistically speaking, is majority of information in there is really high quality information. And so it is the case that sometimes we're going to get it wrong. Most places get something wrong from time to time. It's a good opportunity to think about how you evaluate the quality of information and teach folks how to do some fact checking and how to engage in critical thinking. But overall, if you're looking up the population of a city or a major sort of historical fact, Wikipedia is going to be correct on these issues. And really, if you're going to be making a life decision based on Wikipedia, well, then we'd ask you to consult your doctor or your lawyer or perhaps look for a second opinion.

And if she says, well, explain to me how it truly really works.

You can't just say anything on Wikipedia. There are a couple core policies that require you to cite your information back to a reliable source. So you have to have a citation. You may have seen those citation needed tags. That source has to be what Wikipedians think of as reliable, which means that it has to be respected in its field. So if it's a media publication like a national newspaper, or even a regional or local newspaper, or a medical journal or some sort of other academic journal, all of the information has to come from that. And if it's disputed, then Wikipedians will pause, they will ask for additional citations or additional sources before something can stay on Wikipedia. So generally speaking, the more controversial or the more highly viewed an article is, the more accurate it actually is. In order to contribute to Wikipedia, people need to learn these policies. So yes, it's the encyclopedia that anyone can edit, but in reality, most of the people who contribute to it are pretty experienced in doing this kind of work and work really hard to make sure that the quality of contributions stay high and accurate so that everyone can keep trusting Wikipedia.

What would happen if anybody went to the Planned Parenthood entry at Wikipedia and said, Planned Parenthood sells baby parts? How long would that entry last?

That would not last long. In fact, you probably couldn't make that edit because Planned Parenthood, along with many hundreds of other articles, my guess, if I were to look it up right now, is probably what is called a protected or semi protected article, which would mean that you have to be an experienced Wikipedia editor in order to edit that article. So it's what we think of it as. There's a little lock icon on it, and it means that you have to have advanced permission rights in order to make changes. And that's how we help safeguard articles that are either long term sites of controversy or debate or breaking news events. Information that is sensitive around everything from Pandemics to elections.

So it is literally not true that anybody can edit anything.

Anybody can edit anything if they learn. So I like to think of this as in the pandemic, you may remember, lots of people wanted to foster dogs. In order to foster a dog from a local shelter or an adoption agency, you had to apply to foster a dog. You can't just walk right in off the street and take a dog home with you. There's a level of screening that goes into even this really simple act that hundreds of millions of people do all the time around the world. With Wikipedia, everyone can edit, anyone can edit, but you do kind of have to go through the process of learning first. And in order to make a contribution to an important article, your peer Wikipedia editors have to trust that you're going to make a contribution. That is what we call good faith. That is to say that it's going to make the encyclopedia better and not worse.

This is the last question I have about editing. I face this myself. So in my Wikipedia entry, there are places that I would make corrections or additions, but I can't even edit my own entry because I can't cite a source, because no source may not, because no source may know this fact. So what do I do?

So, generally speaking, Wikipedia is one of the first things we'll acknowledge is the more well known you are, the less complete your biography will feel. Because you may be notable for your professional accomplishments, but perhaps Wikipedia doesn't reflect the fact that you're actually an excellent cook and a wonderful friend and a caring father. And that's probably because maybe you're fortunate enough that someone's written that about you in the past in a magazine article or the like. But sometimes all of the totality of who we are is not fully reflected. And so for me, for example, Wikipedia thinks that I speak more languages than I do because I studied them at some point, but it's never given me a proficiency test. So it over represents some of the details of my life and perhaps underrepresents other things that are true about me that I care about simply because some information has been published and some hasn't. For people who have articles on Wikipedia, we encourage you not to write and contribute to your article yourself, because Wikipedia's other really important policy, in addition to verifiability, is neutrality. And it's very difficult to be neutral about yourself. And so when you edit Wikipedia, editors usually can tell that it's you trying to write about yourself, and they generally politely ask you not to. But if you keep doing it they might block you. So. So that's not a great experience. We tell people not to do it.

I think that one of your great accomplishments is that you have increased the amount of contributors. So how have you done that?

Yeah, hard work. So overall, when I joined Wikipedia, the number of editors perhaps per month had been in a slow decline over the course of the last few years prior to me starting at the foundation. And it's always hard at the scale of something like Wikipedia to isolate exactly what the issues are. You sometimes don't know what you've done right and you sometimes don't know what you've done wrong. And sometimes you can do lots of things and perhaps you've prevented something from being worse and you'll never know. In our case, it's. There were a few really baseline, sort of table stakes things that we needed to do. We needed to make the editing experience more pleasant and seamless and up to sort of best in class experiences people have when they're in other document production environments, whether it's Google Docs or something else. We did that. And now if you're contributing on a desktop, it's a great place, super easy, very reliable experience, very few bugs, really consistent, you can save your work very easily, all these really basic things. The other thing that I think is really important is we launched a mobile contribution experience that was a native experience to mobile, high functionality, high feature set, which when I joined the foundation, we didn't have the ability to really edit on mobile at all. Maybe you could log in through the desktop experience, but it was super kludgy and difficult to be able to do any of the advanced functions. And so we've rolled out this mobile editing experience that is just a dream if you're going along. Remember when we used to ride around on like subways and public transit? Well, I used to commute to the office and I would read a Wikipedia article and I'd find something that maybe wasn't right. And all of a sudden I can just fix it right there on my phone. So that opened up a whole new set of potential contributors to Wikipedia who are mobile first. And so that's the younger generation, people who have more free time that they spend on their mobile phones. Those are people in emerging economies who are mobile first in terms of their contributions. And we saw a tremendous increase in people's usership of these features such as mobile editing. The last thing that I think is really important is, is we have made a number of changes that have been designed to address the culture of contribution to Wikipedia. And what I mean by this is we want people to be friendly and welcoming and encouraging of new editors. And so we rolled out a code of conduct, universal code of conduct for all of our sites that would remind people that it's all about raising the floor of what the feeling should be when you're on Wikipedia so that newcomers feel supported and understand how to best contribute and why and how their contributions are going to be valued and recognized.

What percentage do you think today is mobile versus desktop?

Ooh, I'd have to go back and look at mobile editing. Mobile readership is more than 50% of our readership, and mobile editing was gaining rapidly, but. But they're almost just different use cases. People who write long form articles or on desktop, people who do this sort of copy editing, which is really important, that happens more on mobile, so they're complimentary as opposed to sort of binary one or the other.

And you've also done a great deal of work about gender diversity. So how did that work out?

Yeah, I'm really proud to say that in the last year we received data, even just in this last quarter that indicated that we'd had 30%, excuse me, 50% increase in the percentage of people who identify as women contributors to Wikipedia. So the reason it could be 50% was that it was so abysmally low to begin with. It was at 11% and now it's about 15%. Focusing on the contributions of women, and not just women, but really just any group that has been marginalized or excluded from Wikipedia has been a big priority for me on a personal level and then for the organization as a whole, because we have a mission to support the world with all the world's knowledge. And all the world's knowledge means that it can't just be knowledge that men care about or knowledge that folks in the United States care about, or knowledge that English speakers care about, because the world is very large and vast and rich. And so it's important. It represents and reflects the totality of the human experience. So women are a big part of that, right? Like 50% of it. A little more than that even. And so we've been very focused on increasing women's contributions. Not because they only write about women or only write about things that women care about, but the expectation is, over time that will ensure that we have better representation of all the things that all people care about. About. We have done a number of things to address this. I mentioned the universal code of conduct, which is really around ending harassment on Wikipedia. Harassment is One of the biggest deterrents for women's participation online. On any platform that you go to on Wikipedia, harassment takes a very different form than on some social platforms. It tends to be less overt and more about creating. Creating an environment in which people feel unwelcome. And so we wanted to be very clear about what does welcoming behavior look like on Wikipedia and what does harassment look like and how do we ensure that it's not just about protecting people on our platforms, but also ensuring that bad behavior doesn't follow them off platform to other places where they may make their home online, for example, social media sites and the like. This is really about creating an environment in which everyone feels welcome and everyone feels as though they can participate. And so that was a huge piece of work, Guy. You might remember that we started working on that together in 2016 and now in 2021. That is a fully enabled policy for all of the Wikimedia sites. And so really pleased to see that have come through. There's two other areas that I think we've made significant progress for support for women. One is the Wikimedia foundation is a grant making organization. So we fund efforts that enable and support free knowledge around the world. So we have gone ahead and invested very heavily in efforts by groups of women, non binary folks, anybody who's really focused on gender discrepancies or other racial and ethnic discrepancies around representation to enable them to have the support and resources they need to create partnerships and around free knowledge to build community around free knowledge to do public outreach around contributions to Wikipedia. And we've seen really good success here. We've seen that women organizers of events tend to. Women are overrepresented in those areas of organizing. They tend to engage and stick around longer and become community leaders the more support that they have over time. And so that's a model that we're working to develop out and ensure has the resources it needs to be successful. And then repairing all of this work with efforts within our actual editing interface to surface some of the things that make Wikipedia feel like a more social and community oriented site. Because our research indicates that when women editors in particular know how their contributions are valued, when they have peer mentors that they can speak to in order how to learn the process of contribution, they tend to stick around longer and become more productive and more satisfied editors.

Over time, What happens to someone who violates the universal code of conduct?

So, great question. Right now there are a number of different ways we're experimenting with how to best enforce this code of conduct. Because Wikipedia is so large, we have realized that what works for us is probably very different than many other social sites. So in most places on the Internet, you have hired trust and safety professionals who review indications of harassment or unfriendly or unwelcoming behavior. We're trying to work with our global community to build their capacity to identify problematic behavior and then apply solutions so that it's something that scales to the size of the community, as opposed to trying to make our relatively small nonprofit grow to scale to the size of the world. Wikipedia has always had avenues of enforcement. You can block a user, you can ban a user over time. These are pretty intense sort of interventions, though, and they don't necessarily help people learn how to work more successfully together. And so what we're focused on is building capacity within the editing community. We've got editors who are learning about how to identify what unfriendly behavior looks like. We are developing skills within our editing community around would be called allyship. So how to step in and intervene if they see a pattern of unpleasant or unwelcoming behavior. And then, of course, having other ways to be able to sanction a user in have an appeals process so that if a user is sanctioned, there's always a way to be able to understand the thinking behind it. And of course, if we're truly sort of problematic users who continue to violate our code of conduct, the global permanent ban remains an option.

And I didn't even know such a thing existed until a few days ago when did. Does the Supreme Court kick in?

Yes. So you are referring to the arbitration committee of Wikipedia, and there's one in a number of languages. Yes, those. Those kick in when somebody makes a request that the arbitration committee hears their case. So it's not a default for every sort of incident that occurs. But a Wikipedia editor who disagrees with another Wikipedia editor over a matter of policy, not content, can go to this arbitration committee, which is a volunteer committee of experienced Wikipedia editors, and ask that their case be reviewed. The challenge with this is that it's a very public process, and so it's not appropriate always for cases of harassment, because sometimes that requires hearing evidence in public in a way that can really be a deterrent from someone bringing their case to that court, as it were. And so we've been working with that committee to develop private processes that would enable people to be able to have their cases heard in a way that both is true to Wikipedia's commitment to transparency and accountability, but also enables people who are victims of harassment to not feel like they have to unearth all of that in public in order to be able to have their issues addressed.

Perhaps you can help the U.S. supreme Court in that regard, but we won't go down that path. So this is 2016 was when I was there, 2021, you've gone from 11% to 15%. Why is it still 85%? Why isn't it 50? 50 or 51? 49.

Yeah. So Megan Smith, who's the former US CTO, was at Google for a long time, is very invested in women in technology and thinking about the ways in which to support women in the full sort of understanding of stem has a really nice list of issues that deter women from participating in tech. And it tends to look at issues around do people feel welcome? Do they understand how their contributions are going to be valued? Do they feel as though there is value in contributing? So is there a net personal benefit, whether intrinsic or extrinsic, for that contribution? Those are some of the ways in which women enter into technology and stay into technology. In the case of Wikipedia, there's a number of other factors that happen before you even get to our sites. So when we look around the world, globally, women have less leisure time, which means that they have less free time to participate in hobbies that are sitting perhaps alone with your computer contributing to Wikipedia. They often have not just paid jobs, but sort of jobs around the home. Most housework, child rearing falls on women. The other pieces of this include the fact that women are often deterred from participating in knowledge in a way that is about the construction of knowledge. So while women are educated, at least in the United States, at higher rates than men, what we often see around knowledge seeking behaviors on Wikipedia is women come to Wikipedia less frequently than men do. So before we even get to a question of why aren't women clicking that edit button, we have to understand why are women not seeking knowledge? And is that because we don't have information that's relevant to them? Or is it because there's an actual pattern that deters women from information seeking behaviors that is rooted in other sort of cultural questions about who. Who owns knowledge and who gets to be an authority in our society? So what are the issues that are off Wikipedia which are around how are women accultured to understand their value, their worth, their voice in society as a whole? And for all of those great stats on women graduating with more bachelor's degrees than men, we don't see that consistent through PhD and advanced academic programs into tenureship Particularly in the areas of technology and the sciences. So what are the sort of broader social contexts then thinking about what is the time that women have to participate in hobbies, including contributing to Wikipedia? And how does something like Wikipedia, which is quite abstract, you don't get to see the immediate effects of it in your community, in your kids schooling and the like. How do we make that more tangible for people so that they feel as though there's real value there? And then how do we make our platforms and sites feel as though they are welcoming and easy to contribute to in a way that once people do start they're like, oh, this is great and I really want to stay a part of it.

If you control for all the, the busyness and other things that might remove women from the use of information, much less contribution. I think you said an interesting thing that women even use Wikipedia less than men. Do you have any theories taking off busyness and all that? Is there anything else that why women would not seek information as much as men?

So there are a lot and they, there are sort of a number of things that we're trying to understand about this because we commissioned this research globally and we found that even in societies that are considered incredibly egalitarian, like the skin Scandinavian nations, for example, women still visit Wikipedia less frequently throughout the course of the week than men visit Wikipedia. So we can control for leisure time, for example, as you've rightly said. But there's a question of sort of acculturation around knowledge seeking and the ways in which perhaps, maybe we don't have the right type of knowledge that women are looking for. That's a potential product market fit issue. We see that in emerging markets where we know when we talk to African Wikipedians. We have a real challenge here where many Africans are looking for information about their local pop stars and you don't have it, so you guys just aren't that relevant. So maybe there's a product market fit there. That's true for women as well. But what we actually suspect is that it's something a little bit more pernicious, which is that in general, following formal education, women tend to drop out of the information production space, space outside of formal workplace environments. And so within the context of when a woman is employed as a professional, producing knowledge producing and contributing can be very successful. The idea of engaging in sort of long form individual knowledge seeking and production is something that we just see a tremendous divergence where men are bloggers and sort of substack writers and the like, and you don't see that same representation as an overall gender balance in those spaces. There's also this really interesting thing because when I say this, people go, but what about like Pinterest and Facebook? Women are on these sites, right? And the answer is absolutely. But very often that is a form of communal production and social labor whereby women take on the majority of the responsibilities in their household or in their community of maintaining social ties. And, and those platforms are very productive around performing communal or domestic relationship labor. I'm going to make sure that the kids, I'm friends with the kids, friends, parents, or I'm going to use my Pinterest board to be able to organize what do I need for this upcoming birthday party for what's going on in my community or my family. So these are forms of unrecognized labor that are really important that women perform online that are distinct from the ways in which men perform productive labor online or social labor online. So it, there's just really interesting sort of theories that underlie all of this and unfortunately they're all horrifically gendered because that's kind of the world that we live in.

You're trying to get corporations to pay for some of the information that they get from Wikipedia. So can you explain that?

Yes, we're asking people to pay for a thing, we give them for free.

What a concept. Yeah.

We are our own greatest competitor. So one of the things that I think is fascinating about Wikipedia is that most of us know it as first and foremost the encyclopedia itself. You click on the article page, you read the article, or you skim for the information you're looking for. And in that sense we're really an end user consumer product. But over the course of the last 10 years, we've seen new emergent use cases for Wikipedia that are more backend what I think of as knowledge as a service functions that look at Wikipedia as a database for computational science research, natural language processing training, databases, machine translation, taxonomy building. So there's that piece of it and then there's the reuse of Wikipedia, which is if I have my voice assistant or if I'm searching for information through a search engine, where is that that sort of content forward result coming from? So. So over the course of the last decade, search engines have gone from simple snippets to much more complex results. Voice assistants now have a whole category of utilization that didn't exist 10 years ago. Wikipedia is massively embedded in these experiences that are provided by these other large platforms that you're more familiar with in sort of the consumer space. So the Amazons, the Apples, the Googles of the world, we recognize this as a good because it means that free knowledge is getting out to more people. So if you're sitting there doing your homework and you're like, hey Alexa, I've got this question and you get Wikipedia's knowledge, fantastic. We're thrilled. Same thing is true of search results. But there are some challenges around this as well, which is that it means that people aren't coming to our sites, which means that they're not necessarily editing or donating. We are still a nonprofit. It means that if information is cached on one of these companies servers and redistributed and that information is out of date, you're getting bad information or inaccurate information. And we don't want that either. So we have been working with these larger tech companies to say, what would it look like for you to have great high uptime access to our APIs at an enterprise level where you can have the most up to date information from Wikipedia? If you have data on your side about where something might be inaccurate or wrong, you have a dedicated way to get that back to us. And what would it mean if you, what would it mean if you paid for it? Because one of the things that we think is true is that these companies are getting real value out of the product that we offer. If we're building an enterprise level API, we're going to have to maintain that services uptime time and the like. And so we think it's just fair to say this is a transaction that benefits everyone. It gets more free knowledge out there, you know, the quality and the assurance on our side of things. And we're going to use that money to be able to support the expansion of free knowledge and the, in the sustaining of the high quality of free knowledge. So we're going to keep doing outreach and work with women, we're going to keep growing into emerging markets and languages that are going to be valuable to you at some point in time. We're going to sustain the commons and invest in the growth of that sort of free knowledge base which is ultimately going to have a net benefit to you. Because that's the information that people are looking for, it has value to your users. And so this is a win win all around.

And have they said okay, we are.

Still in conversations, but yes, the indications are positive that this is a service that people really would like to use use.

Because if they don't say okay, they are morally reprehensible in my opinion. But we don't need to go down that path. I am more interested actually in the Wikipedia community's reaction to that because I would guess that there are some people saying that we are selling out to big business.

Of course, definitely there have been questions on both sides or not both sides. With Wikipedia, there's no both sides, there's a thousand sides. Yeah, so. So I think that this is all about how you do it. Right. So it is for us, it's not that we will never sign like a non compete. This is a service that's available to everyone. You can still get the product for free if you want. There are nonprofit, educational free use exemptions. So if you're a university, if you're researchers, if you're the Internet Archive, we want you to have that. Like that's just part of our mission. This is really around an elective opportunity for folks to buy access to a service with a dedicated sort of relationship person that you can contact at any point in time in the day. We'll keep you up to date on the architectural changes that we're making, when we're doing, making adjustments to the service as a whole so that folks can plan for that with their engineering teams. All of that good stuff. We're going to do that anyway. This is just a more formalized way of doing it. And I think that abiding by some of those sort of core commitments around transparency, around there is still a free alternative around this idea that this is a service that offers a net benefit to a certain type of reuser is something that helps us navigate what is honestly a challenging landscape. You know, we want to be able to be as sustainable a product as possible possible over the long run. We want to be able to build out Wikipedia in a way that it is more resilient, more enterprise friendly. So that when over the course of the next decade, as interfaces, experiences, end user needs change, Wikipedia is changing too. From a sort of static HTML based website product into what I really think of as a corpus of knowledge, a enormous database of knowledge and insight that, that frankly we don't make the most of right now. And so how can we invest in our own underlying platform and architecture in order for everyone to make the most of Wikipedia in the long run in a way that is sustainable so that we can continue to thrive and people can continue to have access. When you start to talk about it like this with the community and when we start to say, hey, the revenue that's gonna come from this is not meant to go into the coffers of the wikimedia foundation, certainly not into my pocket. It is meant to go back to the community. It will increase the amount of resources that are available to grow our presence in South Asia. It will increase the amount of resources that are available to invest in Wikidata and Wikimedia Commons and other Wikimedia products that are of tremendous value to the world. People are able to say, okay, I can see the benefit in that. Let's just make sure we keep it transparent. Let's make sure that we keep it non exclusive. Let's make sure that we're connected, continuing to have a conversation around how we govern this incredibly important resource for the world, not just as an economic asset, but as something that has a much broader and more resonant public good.

I would make the case that you are one of the most experienced people in the world at managing communities and fake news, real news, semi fake news, those kinds of issues. So if Mark Zuckerberg or Jack Dorsey called you up and said, Katherine, can I have some advice here? What would you do to fight fake news at Twitter or at Facebook? What would you say?

I would say this is a really hard problem that is unfortunately rooted in structural incentives. And those structural incentives are encoded into the platforms in the way they've been built over many years. But hard problems are also exciting challenges. I think that there are a couple of interesting things that are happening in both of those platforms currently around what are ways that Facebook is experimenting with the oversight for Forward and some interesting models of governance. I think that there's a lot more to do and by no means is it a sure success at this point in time. But I'm interested in this idea that I think is very resonant with Pedia, which is how do you bring in differing perspectives and opinions and create more accountability in larger groups of people to decision making. I think that that is a good direction for them to explore from a policy perspective, from a content moderation perspective. And there are probably ways in which that can be applied even to product development questions themselves around. You know, when we think about when Facebook thinks about its groups products or its feed product, one of the things that we often think about is we don't try to build for end user outcomes, we try to build for community based outcomes. And so while I could build a product that keeps you on Wikipedia indefinitely just clicking away from link to link to link, and that might not be the best sort of collective outcome for the product with editing, we could do sort of things similar and keep people editing individually in their own silos, if we built the product in a way that supported that. But what we're constantly trying to do is actually increase the amount of interaction between ideas and between users, because that sort of interaction creates a productive friction that allows us to say, these are our community norms, these are our community standards. This is is how we get to high quality content. This is how we get to outcomes that benefit a broader group of people rather than an individual group of people. Those would be interesting questions, I think, for both Facebook and Twitter to really think about is, as we're building product and as we're building policy, are we building for end user outcomes, which is how our entire revenue model is structured, or are we building for community outcomes, which is looking at overall communal satisfaction, overall sort of net positive social outcome, within a language environment, a market environment, a regional environment, however you care to sort of segment that out.

If I understood what you just said.

I'm sorry, let me paraphrase.

So basically, as I understood it, that Facebook and Twitter, their entire business model is to keep you on site clicking so they can have, you know, more time, more mind space, and obviously more opportunities to advertise to you. Now you're saying that instead of addicting these end users, you should do what's right for the community, which may in fact mean that some things that would be extremely sticky would not make the cut in terms of being on the site.

Yeah.

Is that what you said?

I mean, that's definitely a part of the argument. And I think you've seen that advanced in other places and by other people. I think it's also around one of the questions that keeps coming up around content moderation and therefore fake news is what are the community norms around, what accuracy is around, what misinformation actually is. And those aren't just about, are we keeping people on the site clicking away endlessly? These platforms have actually come out and said, we don't want that. That's not good in the long run for our business model. We don't want to be sort of a cesspit of information online that harms us with our advertisers, it harms us with us overall with our public valuation, et cetera. So I think that's certainly a piece of it. But the other piece of it is how do you actually build a sense of community norms through building community outcomes? And so community norms around what kind of content do we want on this platform can be segmented down to what kind of content do we want on this platform for this group of language speakers, what kind of content do we want on this platform for this interest group? And then how do we build policies that support those norms based on those community expectations? Right now one of the biggest challenges the Internet has is we have. Well, I mean there's lots of challenges, but one of the biggest challenges around content moderation is really this idea that we actually don't have very good normative understandings of what I find acceptable and what you find acceptable. And a lot of the ways in which we historically built those normative understandings around questions of self, speech, around questions of privacy and the like, were over many decades of jurisprudence, many hundreds of years of co civilization evolution. In towns and in villages and in nations, in cultures. We have telescoped that down into 10 years or 15 years or 20 years of social media. And at the same time we have telescoped the normative time frame down. We have expanded, expanded the network so massively that now we are trying to create norms around what we accept and what we like and what we value at the scale of the globe. And that is a really hard thing to do. So what I would be looking at is how do we actually try to bring that back in, into smaller communities that can create the norms around content, moderation, fake news, privacy in those smaller communities and build product that helps create community outcomes as opposed to product that is focused on individual end user outcomes. That's how Wikipedia does it. And it works.

In that theory, is it feasible that there is a community norm for a community that, that doesn't believe in vaccination, that believes the election was stolen, that believes that there's a lot of election fraud, so we shouldn't be giving people water in line? I mean, are you saying that that's an acceptable norm in some communities or that just is not? That's just beyond everything.

I think that that actually becomes an interesting question for a platform norm. So one of the conversations that's in the news right now and on the lips of legislature legislators is this idea of plat liability and this. So CDA 230 is often what you hear, but it's a conversation here in the United States and in Europe under, under a different name. There's this very interesting conversation that's happening right now around what are the rights of platforms to moderate their own content and what are the rights of platforms to set their own policies and what are the responsibilities of platforms to keep up information that they may find distinct. So for example, when we talk about anti vaccine information misinformation, many of the platforms have taken a more aggressive approach to taking down false information about COVID 19 vaccines or treatments. And they are able to do so because the platforms themselves have a right to speech, because that is recognized within US Jurisprudence. That right to speech can completely be congruent with what I was saying earlier. A platform can have a right to speech that says we're going to have community norms around privacy issues or speech issues. So, for example, we don't want nudity in that smaller subset of a community. And the platform can still also have a right to say, we don't want hate speech, we don't want Nazis, we don't want proto fascism, we don't want explicit forms of racism, hopefully racism in general. But most platforms are really focusing on the explicit kind right now. The platforms can have those norms that they create as their standards and say, this is the floor and we don't want content to fall below it. And then anything above that is something that our communities within those platforms can help set as appropriate to the community itself. So if you've got communities of folks in the United States on the same platform as a community of folks in, in Bangladesh, those two communities may decide that they want different types of norm expectations around information sharing, freedom of expression, privacy, and the like. And that can be okay.

Based on what you've seen, do you think that machine learning and AI can set the floor or at least enforce the floor? Or does it take rooms full of people looking at every post and every picture?

Well, neither solution is great right now for lots of reasons. So AI, it's been promising for so long and remains promising, but has yet to live up to the promise. And that's because some of the issues that we face find most challenging are the issues that really involve human discretion and discussion. And so what constitutes something that is outside the bounds of my standards? Maybe something that is inside the bounds of another country's use of the exact same word. Right. The English language enables us to take have different meaning in different places for the exact same word. So it is highly contextual. And often that requires a human in the loop, closed loop systems, and a human in that loop. So AI can be useful as a heuristic, but it is not sufficient as a solution. On the other hand, you have this idea of sort of hundreds of thousands of people screening everything from nudity to offensive language. And that rigorous application of a few baseline policies, or even hundreds of pages of baseline policies, is never going to be consistent and applicable and is going to require systems of appeal and transparency around how those policies are being applied. So AI is great. I'll give you an example where AI is great. AI is great for identifying content related to violent extremism once it has been determined that an image or a video constitutes violent extremism. So you can have essentially the digital watermark equivalent, which enables systems to be able to flag and identify that video or that audio track or that image and take it down. This is used very widely across platforms. The same is true for child abuse images. So child porn, which is a thing that every single platform that I am aware of is very active about, saying doesn't belong, it's illegal. We're taking it down, we're investing and taking it down. This cannot stay on the platform. These sorts of digital watermarks work really well. And AI is the sort of thing that flags it and enables us to review it and take it down right away and ensures that repetitive signatures of that content come down as well so it can't get uploaded. That's fine for that particular type of content. For issues of speech and issues of political discourse and issues of social discourse, AI is just not sufficient to the challenges that we have today to be able to address those conversations.

Let's suppose that the owner of the New York Times or the Washington Post says to you, katherine, I'm amazed at the ability of Wikipedia to get people to voluntarily pay tens of millions of dollars. So now I want you to come over and help me figure out how we change our business model to adapt to the Times. Mr. Salzberger or Mr. Bezos calls you up and says, katherine, I need your advice. Tell me about my business model, you say?

I would say both seem to have been doing very well recently, actually with digital subscriptions. I think that it's a slightly different question than how we've addressed this at Wikipedia, because we are meant to be a general knowledge resource. And so one thing that is of critical importance to us is developing knowledge that's acceptable to as broad an audience as possible. And because of our commitment to neutrality, regardless of my own political leanings, the idea is to have as broad a base of people who want to engage with our product as we possibly can. And that works really well for us, and it works really well for knowledge. And I think it works really well for people who use and read Wikipedia because. Because you do avoid that audience targeting approach. I think both the Times and the Post have taken an approach that is really about developing a demographic that can support their product. And that demographic is necessarily going to be more niche than a mass market audience because the overall Media landscape is now so fragmented that you do need to be much more intentional about audience development. And so for the Times and I grew up in the Tri State area, it's like that stereotypical sort of northeast liberal, now increasingly coastal national approach that has a certain income point because that tends to be people who actually subscribe. And so you're creating a product that is specific to an audience, and I think the Post is doing that as well for a slightly different audience that is more so political in nature. So I think the bigger question is, can you actually have a national media product that is appealing to a truly national demographic? That is a truly representative national demographic? Excuse me. And I don't think we have that at the moment in the media landscape. And I think that that is one of the biggest challenges with regards to polarization and fragmentation today.

Okay, let me ask that question in a little different way.

Sorry. I'm sorry, guy.

Let's say either of those guys call you up and say, catherine, from the outside looking in, what mistakes do you think I'm making?

I would say, what is your reader's loyalty to your mission? And how do you make it clear that you're creating value for them every single day? So what we know about Wikipedia is that people get value out of it. It's why they donate. It is useful to them every day, and it is useful on a range of different issues. It is useful when you have a question about Beyonce's discography. It is useful when you want to know what's going on with evolution, volcanic eruption on the other side of the world. It is just a useful in your pocket product. And it doesn't judge you if you have a dumb question. And it, you know, doesn't judge you if you spend hours getting lost on it reading about esoteric stuff. It is just there in a very sort of neutral way. And people love that. They develop a real relationship with that. In the same way that if you were a kid, you went to the library and you could get lost in the stacks wandering around. That is the kind of relationship that people have with Wikipedia. Then you layer onto that the mission and the sort of the fact that we don't sell your data and the fact that we don't sell you ads and the fact that we are here for the public and the fact that we're very transparent about when we get it wrong. And the sort of product has a humility and an accessibility to it that I think is very different from other large Internet products today or platforms today. And so you have A really deep relationship between donors and Wikipedia because they value it so much, because it's been there for them in a pinch, because it's never laughed at them for asking a dumb question. The thing that I would want to know if I were a major publisher is how am I creating value for people every single day over time and building a relationship that is meant to last past the subscription cycle.

When you look at your incredible run at Wikipedia and Wikipedia truly is a phenomenon, what's the big lesson?

Such a good question. I'm going to try this. It's not a product lesson for me. It's actually a lesson that speaks to sort of an essential goodness and human character, which is if you give people a reason to do something together and do it well, they will and can. And I think to me spoke to this fundamental generosity that many of us have and a spirit of curiosity that all of us have. So when I think about what makes us the beautiful parts of our humanity, and there's terrible parts of our humanity, but the beautiful parts of our humanity is that we are fundamentally curious beings. And Wikipedia taps into that curiosity in a way that creates a benefit for everyone. Because I get curious about trains and I research them and I write an article, or you get curious about some sort of farming and you research it and you write an article, and then we collectively win. And so I think building these sort of systems that tap into things that are innate to our humanity and enable us to then have something as a collective good is possible. It's not easy, but it is possible. And Wikipedia is evidence of that.

What's the biggest threat to Wikipedia?

Complacency, the idea that it's always been there and so therefore it always will be, is the biggest threat. We operate in the most highly capitalized, highly competitive sector that has perhaps ever existed. That is a threat whether we realize it or not. We don't have a direct competitor, but if we don't keep changing our product, our stack, our offering, our community, all of it, we will eventually be less relevant as the world keeps changing around us. And so I think complacency is the biggest threat. I like to say that if we're not out there in the top 10, then we're just a bunch of encyclopedia enthusiasts. And the reality is that we have to be just as good and just as ruthlessly efficient and just as perfectionist in the products that we build and ship, and just as thoughtful in the policy decisions and advocacy that we engage in as anybody else out there. But we have to do it on a fraction of the resources. And we have to do it faster and better and we have to do it in public. And so it's really hard. And so it's easy to be complacent because it's easier to be, to not pull your head out of the sand and look at all these challenges. But we have to be moving fast because the world is moving fast around this and we owe it to the world for our future and for what, what we offer.

My last question for you is how does Catherine, who ran Wikipedia, how does she do her best and deepest thinking?

It's usually when I'm not actually working. It's when I'm able to go and take a few days off and go to a place where there's no Internet and deconnect for long enough to be curious. And when I can create that step away from sort of exhaustion or the day to day of the zoom calls, to be able to have silence, I can be curious about the ideas that I've been exposed to and create the connections that create the next insight. And so I am a really big advocate that everyone should take down time because that's when we, that's when we do our best thinking.

Have you figured out what's in the future for you?

Not yet. I am have been so deeply in the wiki world with so much of my heart that I need to step away to think about what comes next. But there's so much out there. I'm excited to take my own advice and maybe go for a walk in the woods and figure out the thing that is most exciting to me right now.

Would you cross over into the for profit world or are you just a not for profit kind of gal for the right mission?

I think so. I think that there are challenges that can be solved in both spaces and I would be really interested in working on some of them with more resources, you know. But I think it would all depend on what the challenge was.

What's a challenge you would, you would. What's a challenge you would enjoy facing in the for profit world?

I'd love to work on anything related to climate and sustainability from a market driven approach. I think that there's so much to do there and that the tech is there and the sort of product market fit might not be. The policy space is not where it needs to be. But that is climate is one of the most compelling and challenging issues of our time. It's the existential issue. So that would be something that I think that actually only the market can solve.

If Elon Musk subscribes to my podcast and calls you.

I'll let you know. I'll say thanks.

Deal on a Tesla? Yeah.

I want one person.

All right, Catherine, this has been so lovely. Thank you so much for doing this. I hope you have a greater appreciation for the power, complexity, and quality of Wikimedia. You should not hesitate to use it as a source of information. Kudos to Kathryn, the employees of the Wikimedia foundation, and the over 200,000 Wikimedia editors and contributors. The world would not be the same without the Wikimedia Foundation. I'm Guy Kawasaki and this is Remarkable People. Kudos and gratitude to Jeff C. And Peg Fitzpatrick for producing another remarkable episode of Remarkable People. This podcast would not be the same without the two of them. Until next time, mahalo and aloha. This is remarkable people.