Host Lisa Kiefer talks with Ed Bice, co-founder and CEO of MEEDAN, a San Francisco company building digital tools and programs that promote collaborative verification, annotation and translation for global journalism in the fight against 'fake news.'
TRANSCRIPT
Speaker 1:Your listening to method to the madness or weekly public affairs show on k a l x Berkeley Celebrating Bay area innovators. I'm your host, Lisa Keifer. And today I'm interviewing Ed Bice, cofounder and CEO of media, San Francisco Company building digital tools that assist global journalists in the battle against fake news. You're on the front lines of what is kind of a hostile [00:00:30] environment to journalists right now. Yup. Let's talk about what technologies you're enabling to help journalists out there.
Speaker 2:Journalism has been embattled for a long, long time. The shift we've seen in the journalism threat model in the last five years is we went from worrying about where revenues are coming from. We solved that issue in part by thinking about new commissioned content models and, and then suddenly we woke [00:01:00] up with a new president, this crazy lunatic in the White House. And we looked inward as journalism and journalism tech community and, and we noticed that, that we had lost trust and we'd lost our ability to assert a set of facts and have those prove more durable and influential than a set of provably not facts. And I think that we went from this deep despair over [00:01:30] not having a really good revenue business model to a more existential threat of really not having the words that were writing and the, the stories that we're publishing have influence and have meaning.
Speaker 2:And this is, this is a deeper crisis than, than uh, the business model. Is this when you founded or cofounded meet and, no, we have to go way back. Median has been around since [00:02:00] very early days of social web. It started with the war in Iraq protest and I'm sure many of your listeners where we're at, the protests in the bay area, on the first day that we started dropping bombs, it was a profound global moment. It was the first time I f I felt globally networked even though I'd been on the Internet since it wasn't Internet. There was on that day, this awareness that hundreds of marches were happening around [00:02:30] the world, literally tens of millions of people were taking to the streets to say, this is not what we want. This is not how we should respond. Second Year of the post nine 11 era feels so naive now.
Speaker 2:But I remember thinking as I was walking the streets of San Francisco that wow, this is what it looks like when we're able to change influence history. And, and there was really a sense that the power of this many people could do it. I went [00:03:00] with, um, my good friend Rouge Giuseppe, uh, who is, he was a human rights photographer who had worked in El Salvador Rouge and I were kind of separated and there were some people blocking market street and I was standing on the sidewalk and I can still kind of imagine the, uh, bald, very tall, very large policemen, uh, reached out, grabbed me from the sidewalk, pulled me into the street and said, you can't be in the street. I'm arresting you. There are good cops. This was a bad [00:03:30] cop. I was not intending to be arrested that day, but I was arrested along with I think 1300 other people, straps around the risks and put on the bus and hauled over to pier 39.
Speaker 2:And I didn't know it at the time. I would've, I would've kissed the guy if I had known how he would have changed my life at the time. I wrote an email from that experience, send it to five friends via, I think it was an AOL online email account. And, uh, [00:04:00] one of those friends who was, uh, an environmental scientist wrote back and his dad had started a tool company. He built it up and he and his wife had inherited some money and he said, ed, I want to publish your, the email that you've sent, you know, in my email, questioned what we're doing post nine 11, you know, with my experience as, as a person who had traveled in the Muslim world and who had had homes open to me and just who had just such a different understanding [00:04:30] of the world. You know, I also had the experience of studying with Paul Wells, stone in college, and, and so I had this latent to activism, right?
Speaker 2:And, and I expressed that into this email, you know, just what, what, what the hell are we doing? This is crazy. We're creating generations of, of misery for, for this sculpture. And they Rakhi people. And he wrote back and said, I want to publish this as a full page run in the New York Times so that people can have a different perspective on what's going on. And I know it sounds [00:05:00] crazy, but this feels to me like this incredibly important moment in history. Within 10 minutes I wrote him back, I'll do this, but it's not my, I'm not gonna publish my words in the New York Times. I'm going to go out and find statements from people in Iraq and people in Palestine. I'm going to put those beside statements from Americans and Israelis and we're going to start this peoples opinion project. We will run this as a full page ad in the New York Times and we will go out and source this content [00:05:30] from around the world and it'll get people thinking it'll be provocative.
Speaker 2:The idea was that we would do this, we would publish it, people would be so moved that they would send us money and we would do this again. And we'd start placing authentic, translated content from around the world, you know, into the New York Times. And, and, and kind of expand then to other papers and presses and, and, and is this the digital New York Times or was this the pace? This was the ink [00:06:00] and paper New York Times. And so in June of 2003 w we ran a full page ad that said in our efforts to bring democracy to the people of the world, we keep forgetting about the people of the world and then had these translated voices below that. And it was very inspiring and it didn't work. We know that in terms of the amount of money that it takes to produce and place and ad that goes out to 3 million Sunday [00:06:30] Times readers and the amount of revenue res we received back from that project, you know, it was an utter failure.
Speaker 2:We had a, a short lived organization called the People's opinion project and did global opinion polling. So, so we showed some of the early trends around global opinion of us post nine 11 and post Iraq invasion. That was pretty profound. We were able to, to show that, you know, our actions had resulted [00:07:00] in this kind of loss of faith or trust in America. You mentioned that you had experienced in the Arab world. Did you live there? What was your I traveled and it wasn't the Arab world actually, although all of our work since then has been, but I traveled through Pakistan, through northern Pakistan and into western China Karakorum Highway. It's incredibly beautiful. The way we were treated there was, it was, um, it was formative. Anyone who's listening to this who hasn't gone [00:07:30] out into the world and traveled, you know, that that was my most important education from that early experience.
Speaker 2:The effort in ethos early on was that the media diet, it leads to these really narrow perceptions, which in turn support ill-advised policy decisions. We wanted to broaden that and we saw the internet as a means of doing that. So everything we did in the early days had an online component. [00:08:00] You know, at the time I was, um, I was designing homes designed background, this like, this is a design problem. How do we diversify the media ecosystem? And the thing that we hit on early, early on was that language was such a fundamentally missing piece that the Internet was even in those days and this pre Facebook, but even in those days, the Internet was going to millions and millions and millions of people all around [00:08:30] the world. And it was a bunch of linguistically siloed communities. So no translation, there was no sense. Yeah. Not to speak of.
Speaker 2:And any machine translation was so bad back then that fundamentally useless from that initial failed experiment, I started pounding the pavement, knocking on doors, calling people. And we've got the intention of an Israeli and Palestinian engineer at Carnegie Mellon University's language Technologies Institute [00:09:00] and that put in motion the last 12 years, which, which the patent for that, uh, translation, uh, you with, with, uh, a great technologist too who worked in Senator Leahy's office actually, is this when you founded this is, yes. So 2006. Yeah. Fast forward 2006 and there's, there's a hilarious story of which I think I've never told publicly. Shortly after the, uh, idea, Kinda jelled and the, these guys at Carnegie Mellon are like, yes, let's do this. And, and this [00:09:30] serious linguistic scientist is like, yeah, a crowdsourced human plus machine translation with a reputation model behind it. These are great ideas. We should, we should write this up and for what it's worth, we have a patent on this and still needed some money to do this right.
Speaker 2:So one of my dearest friends in the world, his stepfather's uncle, really, really successful banker in New York. And I asked for a meeting, pretty nervous, but his family had, [00:10:00] had, um, escaped the Holocaust and I knew that he was pretty motivated to contribute back. So I went into his office overlooking Central Park. I had quit my job. I had done crazy things which were unpopular with my family to try to get this thing off the ground. And so I went into his office with very quite nervous and penniless. Uh, I gave him the pitch and I said, you know, language technology plus this thing called the Internet. [00:10:30] Imagine that must've appealed to him because you're getting at the truth. Yeah. Yeah. He looked at me and he pointed at this picture of his grandfather at the end of the conference table. He said, Mr Bias, my, my grandfather is smiling down on you today in 1904 or something like this.
Speaker 2:A Swedish dentist walked into my grandfather's office and dressed in with a vision for language impacting global peace. And I looked at him and I said, Esperanto, [00:11:00] I was going to say that, yes, that came out of the same kind of divisiveness. And he said, precisely, Huh? My grandfather funded Esperanto and I'm going to give you some seed funding to try to put this idea in. Motion. Language is such a fundamental divider. It's not a surprise that many people have said, oh, if only we could talk a common language, you know, the world would be a better place. So that put in motion, meet Dan and MacArthur Foundation was one of those friends. The real break for [00:11:30] us came when IBM put two of their research labs at our disposal. You know, we've had partnerships with IBM and now have a good partnership with Google, Google News Lab, you know, their interest is in seeing more data. And so IBM's interest was in us using this network to bring in more human data on top of the machine processing so that they can improve their models with Google news lab. Now we're looking at how we bring in more credibility, how we can get journalists writing indicators of a, [00:12:00] an article's credibility.
Speaker 1:If you're just tuning in, you're listening to method to the madness, a weekly public affairs show on k a l x Berkeley Celebrating Bay area innovators. Today I'm interviewing Ed Bice, the cofounder and CEO of me, Dan, a San Francisco Company building digital tools and programs that promote collaborative verification and rotation and translation supporting journalists around the world. [00:12:30] So are you a nonprofit? Yeah, so we [inaudible]
Speaker 2:uh, three years ago we were offered a big contract to do software development with one of the large social networks and so we, we did form a for profit. Then and the nonprofit has an equity stake in that. We are a unique hybrid. 98% of our work is with the nonprofit. Now I first heard about, or read about me Dan with the Arab spring, all the protests and you all were pretty instrumental [00:13:00] in translation. Yeah. Yeah. So as soon as we started meeting in, we hired a small team in Cairo and started working on translating Arab media and, and commentary on that media and putting that alongside a US versions of that media or western English language versions of that media and translating the English language into Arabic. So we, we built kind of the Internet's first bilingual side by side news site. We had roots in Cairo.
Speaker 2:[00:13:30] Some of our good friends were in the middle of the the revolution and, and still are, um, we still have deep connections with Kairos, still have employees there. We spun off a media project there two years ago that was just blocked by the state of Egypt when we were doing translation work during Arab spring. The stuff that was coming off of Twitter and Youtube and Facebook was incredibly important. We found that having no way to provide [00:14:00] notes about the sources of that content, uh, no way to really do investigations into the, the assertions made in that media. We felt like that was an area that deserved some development. So we went to, um, some of our funders and said, hey, why don't you help us next phase of media and it's going to be about not just translation on top of the social content, but also a verification and annotation, [00:14:30] annotation, building context, helping and you have specific products for that.
Speaker 2:I was reading about. Yeah. So, so check is, is that product that came out of bar experiences and, and it's, and it's intended to be really simple. It's a tool that allows for collaborative verification notes. It also performs some machine processes, like makes it easy for a journalist to go out and look at the reverse image search. Uh, [00:15:00] so if, if a piece of social media contains an image, uh, we provide a quick link that says, okay, here's where that image has appeared elsewhere. So if you see that it actually came from 10 years ago in Sudan in is not a picture of a current protest in Egypt. Say you've saved yourself a a an embarrassing moment because we are kind of early to that verification space. Google news lab came to us three years ago when they were starting the first draft initiative. So we are one [00:15:30] of handful of NGOs and media orgs that came together to form the first draft group, which is been doing amazing research trainings, kind of leading a lot of the important work in news verification and fact checking space.
Speaker 2:Um, and it's run by a brilliant woman named Claire Wartell who was a before that or the research director at the Tau Center, the Columbia School of journalism and, and first draft is on a, uh, [00:16:00] a steep growth curve. And, and so I think you'll continue to see a lot of really great things coming out of that organization. And I think our contributions to that have been one of the really big success stories out of me. Dan, let's talk about election land, which is an amazing moment in journalism history. The election land project was, um, spun out of that same first draft, Google news lab, me, Dan, but, uh, with the Google trends team and Propublica [00:16:30] in the lead. So propublica really, really loved prep. Yeah. They're amazing. Really great people. So 94 days before the election, I got invited to Washington DC to meet with Scott Klein and from propublica and, uh, Simon Rogers from Google trends and, uh, Clair from first draft and small set of people.
Speaker 2:And they're like, well, we want to do nationwide election monitoring, you know, with a thousand journalists [00:17:00] 90 days from now. Uh, yeah. And at this point we were in, in a rewrite of, of our software. And so I said, yes, of course. So it was, it was a mad dash to pull that project together. And it's now collecting all sorts of awards. There's now a case study, uh, we've recreated this, uh, for the French election now with a project called crosscheck and a UK general election project as well with a popup newsroom component that had [00:17:30] a bunch of journalism school students together. So the model is evolving. Election monitoring has historically just been this, you know, big agencies checking boxes and observing things. And so this is really the recognition that the Internet, the social media landscape is this incredibly valuable area to do election monitoring to understand how elections are, are working in real time and try to respond to that.
Speaker 2:So, so I think [00:18:00] there's something really important in this. The outcome was that you discovered there was no election fraud. Yeah, yeah. Shortly after the election, Trump was saying there was fraud. Yeah. So has he not seen this data now? He, um, you know, the, the, the irony is that w, you know, 94 days before the election, Trump had not talked about vote rigging. So we're, we're starting this project and we were like, Yep, you know, we're going to be in an amazing [00:18:30] position to look at voter day issues. And then, I don't know, 30 days, 40 days later, Trump says the vote's going to be rigged. And we're like, oh my God, he's just, he's doing our advertising, you know, marketing this project and, and, and making it incredibly important. But there were hundreds of articles that were filed from the findings on election day.
Speaker 2:So the, the model that we had was, uh, work with a bunch of journalism school students and 300 local [00:19:00] media partners and source these stories in real time. I mean, it was a, it was a remarkable and remarkably complex operation, but we were signaling out to reporters during election day and the result of those signals was, I want to say between two and 300 stories may be off on that in terms of the comprehensive view from, there wasn't voter fraud. Propublica did a series of stories on that. Okay. So that was major, that was a pivotal moment, [00:19:30] but very costly, right? Yeah. Costly. Costly in terms of we had a hundred people in the CUNY, a journalism school newsroom on election nights. A certainly there's costs getting all those people together, but when you think about the person hours, we had a thousand journalists using the software. We had about 700 in check and about a thousand on the slack.
Speaker 2:We use slack as a communication back end for the project. When you think about the person hours [00:20:00] that went into that, that came out of that project was pretty efficient investment. So this will continue. I would be shocked if we didn't do midterms and, and sh I believe election land is a models going forward and I think that first draft and pop up newsroom as global election monitoring efforts and, and the research that comes out of that is gonna. I think we've invented a whole new mode of election monitoring and, and I think it's gonna [00:20:30] be a really powerful and important tool, especially as we see the kind of weaponization, the misinformation campaigns that are now being waged around elections. The Bot armies that are being deployed to just, you know, misinformation. All of this needs to be addressed in, in efforts that identify and call out misinformation, disinformation campaigns in, [00:21:00] in the runup and, and, uh, into election day.
Speaker 2:David Remnick New Yorker, he talks about this as the golden age of ignorance. Yeah. We're in, how do we, uh, fight the media moguls who take over, for instance, the guy who owns national enquirers now trying to take Time magazine. And all of those assorted of magazines and that's editorial content that how, how do we get around that kind of gaming? The answer used to be the Internet, you know the Internet, it'll save us [00:21:30] from, from this. Have you guys all talked about that you, you were just at a conference in Italy International Journalism Festival at a certain level, the same consumer appetite that had people clicking onto bula and Outbrain's ads as a means of supporting serious journalism is now supplanting serious journalism with that sort of reporting in there. There are some good signs in subscription models and [00:22:00] what's happening for the post and the Times.
Speaker 2:There's a lot of people who are saying, Oh okay, we do need to pay for this. And maybe that's one positive trend out of all of this. But the idea that people who are just dead set to promote agenda driven media are, are going to control influential. You know, Fox News feels, feels very innocent in comparison, [00:22:30] you know, with these efforts. So the idea that you know, that Breitbart would become as influential as foxes is, is David Pecker with the time empire? If that happens, it's terrifying. So that's continuing this silo, like people who believe a certain way, they know which outlets to go to and you and I may go to Propublican read what we know to be the truth. Are we never going to have the mediation between these groups through journalism that [00:23:00] that's the hope. But I mean really the, the deepest hope for journalism is that, um, the truth, uh, has more weight than untruth, you know, if that is thrown out then, then the sorting mechanism, just his, because it all is in, is all about the truth.
Speaker 2:It's, it's there. It's supposed to help us. Yeah. Yeah. I, I think that we're in some, some really, really dark days and, and that, [00:23:30] these technologies that we thought were, you know, so liberating and so empowering and the wisdom of the crowd that would, that would surface and, and, and the sort of Wikipedia model re across human knowledge that would have affordances for editing and annotating and revising every object knowledge until it came to the point that was like, was better [00:24:00] or more true as we wade into conversations around the truth. One thing is that working in a global context, you're really humble about this truth. You recognize that there are a thousand truths that describe an important piece of every event. It's not just to descend into total relativism, but to acknowledge that context is, is always dependent on a cultural framing, [00:24:30] the reader framing the understanding the source better.
Speaker 2:So I feel like I want to offer this disclaimer that as, as I'm saying, you know, we need the truth to mean more that I'm not saying there's not just there, there is one truth in the end and you don't have no, nor should there be one arbiter of the truth. And, and right now the one thing is very concerning for journalism is, is the, is the fact that Google and and Facebook are distributing and Twitter distributing, you know, these, these [00:25:00] are distribution pipelines that are so dominant right now, surely in terms of how the search algorithms and newsfeed are influencing what we're getting on a daily basis in our media diet. Those are the platforms that are very, very serious about saying we don't want to be arbiters of truth, but the algorithms that power newsfeed in power search are arbiters of meaning. And that is, is a pretty close proxy for truth.
Speaker 2:You know, I think [00:25:30] we're in some really dark early days of understanding, um, how these systems, uh, were where it a failure point I think there is resolve to try to do better. And that's, that's, that's changed a lot since, uh, early November. They understand the problem and neither one of them is, is saying, oh, we need to build a truth algorithm, which is really good. And our role over the next year is going to be helping think about how signals from journalists [00:26:00] are treated by those platforms. So having a way of looking at how 30 or 40 different journalists from around the world are, are viewing a claim that might be circulating and, and then surfacing that into a Google search result as as a fact check. Would it post an alert to the yes or Google started doing this already. So Google and in some cases if you're on Google news and article contains [00:26:30] a claim that has been fact checked, they're just in this just in the last month starting to surface.
Speaker 2:In fact check Facebook has dispute. We can better structure signals into those types of systems. I really respect the technology building you're doing for journalists. I think it's, it's really important. I worry about the flip side of that. There is less curiosity today because of some of the technologies that have been built. Readers become [00:27:00] lazy. They don't do the deep connecting. They put trust where they shouldn't. What do we do about that? That, I mean it's, the technology is partly to blame for that. Before we had to open an unfolded the newspaper. Yeah, it is nanny's garden next to, you know, bombing in Yemen. I think the response to that is, well, two things, decay of society motivates people to realize that sitting [00:27:30] back and allowing the media system to decay has some real bad consequences. But also thinking about tools that allow people to, to feel that they have, um, more agency than just putting up a, uh, a smiley face or a, you know, a, a sobbing face in those go to structural issues with the web.
Speaker 2:How do we Wikipedia FY the Internet in a better way [00:28:00] so that even citizens can write signals in a structured way that a journalist who wants to take in those signals or who's looking for them or who's maybe gotten a really credible signal from that person before might look at and say, oh my God, this needs to be written about because it's going to change this story. What's coming up for you in the future? Bridge is our translation project. So we're working to bring that product into open [00:28:30] source and we're also looking to integrate bridge as a translation solution with check, which is verification solution. So a lot of these events that break around the world are reported outside our language community, giving journalists a good way to get firsthand data, get that professionally translated, then do verification work. On top of that, we're working with some, some stellar partners.
Speaker 2:So we've got projects in the pipeline now with the Syria of Video Archive, [00:29:00] uh, which is a really important project to archive and mark up videos, uh, that we hope will have evidentiary value. The Digital Verification Corps, which is, uh, Amnesty International and Berkeley Human Rights Center project. Some of those projects are in the pipeline. We are keen to, to c check in more newsrooms in the u s and to repeat the election work that we've, we've, uh, been doing. If somebody wants to get ahold of you or, [00:29:30] or go to your website, if you could share that with me. Yeah, we're at me, Dan. It's m e e d a n.com. And uh, can always send me an email@helloatmedia.com. That email will go to me and my colleague on show, Mina and, and Tom [inaudible]. Anyone who's interested in contributing to open source software development or helping us think through some of these thorny issues that we're working on that we'd love to hear from you right now is this moment in history. [00:30:00] We need technologists, we need journalists. Uh, but we also need philosophers. I think we are dealing with issues of truth and ethics and we, we've created hugely powerful technologies and maybe we've lost our way. Maybe we needed more philosophers and academics involved in thinking through what this would mean.
Speaker 1:Thank you so much and yeah, no, it was, it was my pleasure. That was Ed bice, the Co founder and CEO [00:30:30] of me. Dan, you've been listening to method to the madness, a weekly public affairs show on k a l x Berkeley, celebrating Bay area innovators. You can find all of our podcasts on iTunes university. We'll be back again next Friday at noon.
Hosted on Acast. See acast.com/privacy for more information.
