Latest / Stage Talks with Bellingcat / "How We Got Here: The Rise, Backlash and Future Of Counter-Disinformation" with Eliot Higgins and Dr Claire Wardle
Transcript
- 0:04You're listening to a stage talk titled How We
- 0:08Got Here, The Rise, Backlash and Feature of Counterdisinformation.
- 0:13In this episode, we're joined by Bellingcat founder
- 0:15Elliot Higgins and Cornell University professor
- 0:18Dr. Claire Wardle. Together, they unpacked the
- 0:22historical context of online fact -checking and
- 0:25where open source investigations really began
- 0:28to shine. You can find links to all the resources
- 0:31mentioned in the talk in the podcast description.
- 0:34This talk was hosted by me, Charlotte Ma, on
- 0:36Thursday the 29th of January, 2026, in the Villain
- 0:39Cap Discord server. Hello, beautiful people.
- 0:52We're back in a new year with a new set of talks
- 0:55to kick off our 2026 series. We're discussing
- 0:59the rise, fall and future of counter disinformation
- 1:02and fact -checking efforts. Something central
- 1:04to the work we do at Bellingcat and important
- 1:07to the balance of our society. Someone who's
- 1:10been a leader in the study of the way we receive
- 1:12and respond to information for decades is Dr
- 1:14Clare Wardle. Claire is a co -founder of the
- 1:17former First Draft News, a non -profit providing
- 1:19practical and ethical guidance in how to find,
- 1:22verify and publish content sourced from the social
- 1:24web. She's currently an associate professor at
- 1:27Cornell specializing in information ecosystems
- 1:30with a focus on effective communication in the
- 1:32age of AI and conspiracy. If you're familiar
- 1:35with Bellingcat's work, our other speaker needs
- 1:37no introduction. We're joined by Bellingcat founder
- 1:39Elliot Higgins. Elliot has been working with
- 1:42the University of Nottingham on a new framework
- 1:43demonstrating the importance of epistemic security
- 1:46for a healthy democracy. Today they'll be breaking
- 1:49down the emergence and journey of counter disinformation
- 1:51strategies and answering your questions on what
- 1:54our future looks like as people are now fed information
- 1:57according to political preferences and life experiences.
- 2:00Whilst we talk you can place your questions in
- 2:03the chat accessible on the right hand corner
- 2:05of your screen but please remember that this
- 2:07is being audio recorded for the podcast So if
- 2:10you don't want me to mention your username, please
- 2:13add that as a note in your question. Okay. Elliot,
- 2:16Claire, over to you. Oh, yeah. Thanks for coming
- 2:20to speak to us, Claire. I think I first met you
- 2:24probably 20, my memory's terrible. So 2011, 2012,
- 2:29the very early days of this whole kind of, I
- 2:34guess, movement you could call it. This, correct
- 2:37me if I'm wrong, this started for you when you
- 2:38were in Storyful and maybe you can tell us a
- 2:40bit about what Storyful was and how you kind
- 2:42of came into this area. Yeah, so actually it
- 2:47goes a little bit earlier than that. I was at
- 2:49Cardiff University as a professor and wrote a
- 2:51report for the BBC about user -generated content,
- 2:54because in 2008 they were like, we're getting
- 2:56this stuff from the audience and we don't know
- 2:57what's true. So I was like, oh, interesting.
- 3:00And so I wrote the report thinking that only
- 3:01my mom was going to read it. And one person from
- 3:03the BBC read it and said, oh, would you come
- 3:05for six months and help us implement your recommendations,
- 3:08which never happens as an academic. And while
- 3:11I was there, many of you may remember this, but
- 3:13that plane landed in the Hudson River in January
- 3:152009. And the head of BBC News Gavin called me
- 3:18and said, you're the Twitter lady. None of my
- 3:20journalists are on Twitter. We got scooped by
- 3:22CNN because we couldn't find that image. We couldn't
- 3:24verify it and we didn't know whether we could
- 3:26use it ethically or legally. Can you develop
- 3:28a training course for the BBC on how do we verify
- 3:31and use this kind of content? So I just was like,
- 3:33had this niche area of like, how do you verify
- 3:36what's on this? social web and then did that
- 3:39worked for a couple of years independently and
- 3:41traveled the world training journalists. I mean,
- 3:42I was so lucky and it was an extraordinary time
- 3:45to go to Al Jazeera and to go to ABC Australia
- 3:47and go to BBC Africa and like many countries
- 3:51in Africa to train them in how do you tell whether
- 3:53or not the audience is telling the truth. But
- 3:55then from that in 2012, I moved to Storyful when
- 3:57it was still very small, with an incredibly small
- 4:00group of people trying to figure out the Syrian
- 4:04war. And at the time, I didn't even know your
- 4:05real name, right? Like you were just, they were
- 4:07like, this is Guy, he's in Leicester, he's really
- 4:09good at munitions stuff. And you were like, you
- 4:12know, Moses Brown. And we didn't, we never met,
- 4:14we just knew you. And so there was like this
- 4:16crew at Storyful, this small crew at BBC, anti
- 4:19-Carvin, NPR and you. And we were just trying
- 4:21to figure this stuff out. tell people about that
- 4:26kind of period. It was such a small group of
- 4:27people who are kind of figuring out exactly even
- 4:30what words to use. Open source investigation,
- 4:33it wasn't even a term we were using. Someone
- 4:35started using Austin and that kind of stuck and
- 4:39that's kind of plaguing us a bit now, I think,
- 4:41unfortunately. A good way to spot hoses as well.
- 4:44But also, Words like geolocation, I can't remember
- 4:48where that word came from, but it kind of then
- 4:50become just the word we use to talk about what
- 4:53we now call the process of geolocation. But for
- 4:56me, at the first time, I spot the difference
- 4:57with Google Maps. Yeah. Well, because at story
- 5:01for the time, there was Malachy Brown, who then
- 5:03went on and many people probably know, works
- 5:05for the New York Times. And I remember at the
- 5:07time he wrote a blog post for us, which was like
- 5:09the steps to verify a video. And I think I said
- 5:12to him at the time, I was like, I think this
- 5:13might one day go into a museum, Malachy. Like
- 5:15it was this crazy attempt by one individual to
- 5:18try and lay out the steps that he was doing every
- 5:20day. And, and I think lots of that was everybody
- 5:23was helping each other to be like, well, what
- 5:24are you doing there? And, you know, and I remember
- 5:26the first time that we tried to analyze an explosion,
- 5:29a distance of an explosion. And we were like,
- 5:31Oh, has anybody done this before? Like, how do
- 5:33you figure that out forensically? So yeah, it's
- 5:35kind of extraordinary what was happening at the
- 5:36time. And I don't think we realized how important
- 5:39that time was. Yeah, I think that's kind of your
- 5:42experience when you do this work because everything's
- 5:44so new. You're saying how like measuring the
- 5:46explosion. That's something I remember doing
- 5:48in a completely separate case, measuring the
- 5:50range of a rocket launcher and based on how far
- 5:53away the noise of its rocket explosion was from
- 5:55the origin. So yeah, I mean, and I guess Syria
- 6:01was the kind of dominant topic in 2011 and 2000.
- 6:05probably till about 2013, I would say, especially
- 6:07with the sarin attacks. But eventually you moved
- 6:10on to another organization from Storyful. Yes.
- 6:16Well, I just want to remind, and I had just another
- 6:18moment. I remember one of the early chemical
- 6:20weapons attacks where we all figured out that
- 6:23the different platforms had different timestamps,
- 6:25right? Because Putin had said that it had happened
- 6:28yesterday and we were like, what? And it was
- 6:29because we hadn't figured out that YouTube was
- 6:31Pacific Standard Time. And then we're like, but
- 6:33that's different to Twitter, because Twitter
- 6:34depends on where your settings are versus all
- 6:37of this stuff now that we take for granted. But
- 6:39it's incredible to remember those kind of moments
- 6:41that were so defining. But yeah, after StoryFile,
- 6:44I went back and did a fellowship for a year of
- 6:47updating that user -generated content research.
- 6:49I did that at the TAL Center of Digital Journalism.
- 6:51And then around that time, founded a nonprofit
- 6:54called Eyewitness Media Hub. which was a small
- 6:57group of us trying to figure out like what are
- 6:59the ethical uses of eyewitness media? And I have
- 7:02to say this week there's been some incredible
- 7:03pieces written about eyewitness videos from Minnesota.
- 7:07And obviously that's what we've all been looking
- 7:09at for a long time, but it's kind of amazing
- 7:12to think back to, there were real questions around
- 7:14that time, like the Charlie Hebdo shooting. I
- 7:17don't, people may not remember this, but there
- 7:19was a guy in his apartment and he filmed the
- 7:22killing by the security forces and then uploaded
- 7:26to Facebook, not thinking. Somebody else scraped
- 7:29it from Facebook, uploaded it to YouTube and
- 7:31monetized it. And all the newsrooms were using
- 7:34the YouTube version, not realizing that they
- 7:36should check the Providence, right? And so again,
- 7:38there was such uncertainty in those early days
- 7:40of like, what were guidelines for newsrooms?
- 7:43What were the best? ethical practices. And I
- 7:45think we're seeing some of the conversations
- 7:47around Minnesota. I'm like, Oh, actually, there's
- 7:50a whole host of earlier stuff, because I got
- 7:51interested in this with the 2005 777 bombings.
- 7:55which was the first time the BBC ever led a news
- 7:58broadcast with anything from a camera that wasn't
- 8:00BBC owned. Somebody was down in the tube tunnels
- 8:03and had not an iPhone, because it was 2005, a
- 8:06grainy photo from inside. And it was the first
- 8:08time the BBC had led that, right? So when we
- 8:09think about the tsunami in 2004, 7 -7 bombings,
- 8:14blah, blah, blah, it's kind of a fascinating
- 8:16story of the kind of evolution of eyewitness
- 8:17media. Yeah, and I think as well, I've been thinking
- 8:20about the past 20 years in terms of how our whole
- 8:23information ecosystem has changed. And in parallel
- 8:27to all the kind of really dark stuff that's been
- 8:29happening and obviously that we battle against
- 8:31every day now. There was also this growth of
- 8:34this parallel community that were truth seekers,
- 8:36but trying to figure out actually how do you
- 8:38establish the truth? I remember, I think the
- 8:40big challenge for a lot of newspapers and media
- 8:42organizations were citing people like the gay
- 8:46girl in Damascus blog, which turned out to be
- 8:48a white guy in America. And that really shook
- 8:50their faith in their ability to trust that they
- 8:51can go online. It seems silly now, you shouldn't
- 8:54trust stuff on the internet, but back then all
- 8:56kinds of stuff was being reported on just because
- 8:58it became It seemed legitimate. I do remember
- 9:02one really bizarre one. There's something awful
- 9:04for a music called Martin Random who did a series
- 9:06of posts about the Bush administration talking
- 9:09about Donald Rumsfeld having to wear iced underwear
- 9:12and all kinds of ridiculous things. It got picked
- 9:14up by the Daily Cast and went a bit semi -viral
- 9:16because of it. Just those basic things now that
- 9:19would be completely ridiculous to do just weren't
- 9:21established within journalism and the blogosphere
- 9:24also had a big influence that suddenly we were
- 9:26getting these alternative voices. people didn't
- 9:30know who was kind of the respectable voices very
- 9:33early on. So you got that kind of weirdness emerging
- 9:35from that. Yeah. And of course it wasn't Syria,
- 9:39but Hurricane Sandy, which was the end of October
- 9:412012, was the first time I remember seeing like
- 9:45all sorts of fake, you know, people like, oh
- 9:47my God, look at this, like the Statue of Liberty.
- 9:49And it was from like some disaster movie or,
- 9:52you know, really obvious, or, you know, there
- 9:54were users that were talking about tubes, subways
- 9:58being flooded and they weren't. And I think that
- 10:00when I had this moment of like, hang on, there
- 10:01was so much more misinformation that up until
- 10:04that point I was training newsrooms saying like,
- 10:06well, once in a while you might get hoaxed. And
- 10:08that was the moment I was like, oh no, hang on,
- 10:10this is a completely different level and how
- 10:12in real time a newsroom is meant to verify this
- 10:14kind of content. When do you think was the kind
- 10:17of moment you really saw that sea change? What
- 10:19kind of year would you say that was? So actually,
- 10:23I think even though it sounds silly, but I think
- 10:25it was Hurricane Sandy because I think there
- 10:26was something about Syria, which was like there
- 10:29was nobody else in, like there were no journalists
- 10:31allowed into Syria around that time. So we were
- 10:33only able to rely on YouTube. And so a lot of
- 10:35it was like, can we verify that that bomb did
- 10:38land on that day? Even though that footage is
- 10:41probably from around this time, we were trying
- 10:43to be really specific. I think Hurricane Sandy
- 10:45was like, oh, some people just want to like cause
- 10:47like create content and call shit just because
- 10:49they can. Like that for me was this moment of
- 10:52like, oh, something different is happening here.
- 10:55Like, which I think up until the stuff that we
- 10:56were looking at Syria didn't feel like that at
- 10:58all. I started feeling around probably 2013 is
- 11:02when the social media companies were engaging
- 11:04a lot more with this. I remember, of course,
- 11:06there's the Google Conflict and Connected Worlds
- 11:09Summit in late 2013, where it was the first time
- 11:12I kind of personally had done a major event,
- 11:14which was terrifying for me. And there are lots
- 11:17of people there from like the Syrian kind of
- 11:18community and journalists and It was the first
- 11:21time I'd seen people talk to each other in that
- 11:23way and people caring about that because before
- 11:26that it was just me and a few people on the internet,
- 11:28including yourself, doing this. Where did you
- 11:31start feeling this was something that was being
- 11:33taken more seriously by the social media companies?
- 11:39So it's a good point. I mean, in 2014 I worked
- 11:42for a year at UNHCR as like head of social media
- 11:45and at the time we had real problems with smugglers
- 11:47basically creating content of like beautiful
- 11:51pictures of yachts using UNHCR logo and that
- 11:54was 2014 and we were trying to contact Facebook
- 11:56and others and they were like freedom of expression,
- 11:59everybody has a right to pretend to be a smuggler,
- 12:01maybe smugglers are good. These kind of crazy
- 12:03conversations with these Silicon Valley bros.
- 12:07So I think around that time, they weren't, newsrooms
- 12:10were. When I was doing interviews with newsrooms
- 12:12in 2013, 2014, they were like, we are terrified
- 12:15of getting it wrong. And I'd be like, so are
- 12:17you building up these skills in -house? And they
- 12:19were like, no, we just use Reuters and AP. BBC
- 12:22was building it in -house, but many, many newsrooms
- 12:24were like, we don't know where to start with
- 12:26this kind of content. I think the social media
- 12:28company as well, I think the point that you're
- 12:30probably trying to get to is in, it was 2015,
- 12:35basically Google invited a small group of us
- 12:38to a New York meeting. They didn't pay for our
- 12:41flights. But we all were like, oh, because at
- 12:44the time we were like, oh, Google, exciting.
- 12:45It's like Willy Wonka's factory, right? And we
- 12:47all, you were there. There was people from Storyful,
- 12:50Eyewitness Media Hub, Medan. There were different
- 12:53people there. And they basically said, we want
- 12:55to start a non -profit. And we want to call it
- 12:57First Draft. So I'm sure we're going to get to
- 12:58more of that, but it is really important to remember
- 13:00that First Draft really started with a number
- 13:03of small number of people who'd been doing this
- 13:05work for five or more years. And it was like,
- 13:08well, how can we amplify that kind of work? But
- 13:11it's really important to remember the foundations
- 13:13came from and like Craig Silver was in that group.
- 13:15Like it was this amazing group of people that
- 13:17were thinking about verification and from different
- 13:20perspectives. Yeah. And that was the Google ideas.
- 13:23I think they called it at the time. was part
- 13:25of that as well. They were very keen to support
- 13:27that. I remember a lot of excitement in Google
- 13:29about how their tools were being used for the
- 13:31human rights thing and then ISIS came along and
- 13:33then they got very worried about how their tools
- 13:35were being used for other things as well. And
- 13:38that's when, especially Facebook, I think as
- 13:40well, you've got a lot of concern. Anytime we
- 13:42tried to speak to Facebook in that period, it
- 13:43was like, oh, we'll have to talk to our lawyers
- 13:45about this. It was like, we don't want to engage
- 13:47with it. Google are always a lot kind of more,
- 13:49I think, communicative. I think they still have
- 13:51their, you know, do goods idea, which kind of
- 13:53obviously, Tamiya's laser is a bit different,
- 13:56but. Yeah, so let's talk about First Draft then
- 13:59because I remember meeting people like Ed Weiss
- 14:02from Medan and him trying to do work and all
- 14:06these different news organizations coming along,
- 14:08all these kind of independent and unusual internet
- 14:10ones. But we still didn't have that kind of,
- 14:14I guess we hadn't really written a lot of stuff
- 14:18down that we could show of people. I think that's
- 14:20something that First Draft did a really good
- 14:22job of kind of trying to start off with. Yeah,
- 14:26and at the very beginning, I mean, Google gave
- 14:28us an amount of money. It wasn't that much and
- 14:30said build a website, which was kind of terrifying
- 14:32because nobody, you never want to build a website
- 14:33for Google. But what we decided to do as a group,
- 14:36as a coalition, is that we said every week we'll
- 14:39write one blog post, right? And so you'd write
- 14:41something about something you've done, I'd write
- 14:43something and then Craig Silverman would write
- 14:44something and then me down would write something.
- 14:46And I am a little bit biased, but at some point
- 14:49somebody will go back to those early first draft
- 14:52writings and see it as an archive. When you look
- 14:55at everything that was published on first draft,
- 14:57it's this kind of incredible six -year journey
- 15:00of the ways in which we were all thinking about
- 15:02the problem. There was a lot of stuff about definitions,
- 15:05and then there was a lot of stuff about bots,
- 15:07and then there was a lot of stuff about Russia,
- 15:08and then there was a lot of stuff about monetization,
- 15:09and then there was a lot of stuff about pre -bunking.
- 15:11You can trace the ways in which this field shifted
- 15:15based on what people were thinking about at any
- 15:17given time. It is kind of extraordinary to go
- 15:20back and think about it as an archive of exactly
- 15:22the conversation that we're having today. I think
- 15:25then 2016 was obviously a big moment because
- 15:28of the exit of the Trump election. And Uterte
- 15:32in the Philippines. Bellingcat, I was busy with
- 15:35all my Syria stuff with Russia starting to bomb
- 15:38Syria after we've done the MH17 investigation.
- 15:41I remember around that time, we were also getting
- 15:45invites to various things. I remember this Facebook
- 15:47Paris event where they were trying to convince
- 15:50a bunch of fact checkers to work for basically
- 15:51free to fact check their website for them. I
- 15:54kind of left that meeting thinking I'm not going
- 15:56to do that with Burning Cat. But what kind of
- 15:58thing were you doing around that period in the
- 16:00kind of aftermath of those events? Yeah. So,
- 16:04so I came on board full time by that time, first
- 16:07draft, I think had gone from this kind of like
- 16:08cute hobby. website where we all write one thing
- 16:10a week to hang on, there's something more here
- 16:13and people need training and we thought about
- 16:15training materials. But just after the US election,
- 16:18we heard from French journalists who said, well,
- 16:20we're worried about what might happen in our
- 16:22April election, April 2017. And so a group of
- 16:25people got together in Paris, lots of different
- 16:27journalists and tech companies and other things.
- 16:30And we kind of came up with this idea called
- 16:32CrossCheck, which was that these different newsrooms
- 16:34would work together. The point being, which is
- 16:37the collaboration that we see at Bellingham.
- 16:38at which is nobody should be competing. around
- 16:41clearing up the shit on the internet. There's
- 16:44reasons that journalists compete and it's around
- 16:46quality information and it's about investigations,
- 16:48but debunking and just cleaning up the mess,
- 16:51nobody should be competing about it. So the idea
- 16:53was what if journalists actually worked collaboratively
- 16:56to write one piece that they could all basically
- 16:59syndicate and all publish so they would be using
- 17:01less resources to create the same content. And
- 17:03then you would also have this amplification effect
- 17:05because the same stuff would be appearing. Those
- 17:07of us who study Russian disinformation, that's
- 17:09what works. firehose of propaganda is that you
- 17:12need to be repetitive and actually quality information
- 17:15sources often don't do that, like we are competing.
- 17:18So that was Cross Check France and that happened
- 17:20in 2017. At the end of the election, everyone
- 17:22was like, thanks so much, no more funding. We
- 17:25did another project like that in Brazil in 2018
- 17:27called Comprova, which still continues. I mean,
- 17:29I think it's one of the things that I'm most
- 17:31proud of is that that coalition of major newsrooms
- 17:33in Brazil still works together to counter bad
- 17:37information in Brazil. So like many things, Brazil
- 17:40have got it right and we've got it wrong. But
- 17:42yeah, so first off was ever a fact checking organization,
- 17:45but it was how can we connect the dots and train
- 17:48people and understand that collaborating around
- 17:52these issues is actually a much better model.
- 17:54and competing. And what was lovely is that you
- 17:56would have these different French journalists
- 17:58together. in Slack. I mean, this is another amazing
- 18:01archive because the French are so philosophical.
- 18:02They'd be like, but what is the meaning of truth?
- 18:04Like all in French, like just these great conversations
- 18:06about like what they were doing. Um, but you
- 18:09also saw journalists from different newsrooms
- 18:10were like, oh, you guys do that. Oh, that's interesting.
- 18:12We do that. Like, and it made me realize how
- 18:14rarely do journalists compare notes. And also
- 18:17the quality of the work was higher. This is what
- 18:19we found out after was in the evaluation because
- 18:21it was said we didn't want to screw up in front
- 18:23of our competitors. So the work they were producing
- 18:26was of a higher quality, right? So it was a fascinating
- 18:29way of thinking about collaboration as a model,
- 18:31which I don't think has taken off in the way
- 18:33it should have done. But yeah, it was an exciting
- 18:35time. I think I often found with those projects,
- 18:39again, it's that funding issue. They'll fund
- 18:41you for one project, but beyond that, they don't
- 18:43offer the support and they just aren't interested
- 18:45in sustaining those kinds of projects in the
- 18:48long term. And I think looking at what's happening
- 18:51in America as well, Thinking about how that collaboration
- 18:53is so important, I would like to see a lot more
- 18:55of that happening in the US at the moment. And
- 18:57it's kind of a shame that America, I don't know
- 19:00if they feel that they're kind of special compared
- 19:02to the rest of the world, but you know, we've
- 19:03seen this happen before in a lots of other contexts.
- 19:06We have processes and ideas and we understand
- 19:08why building communities around verification
- 19:10is important. I feel those lessons need to be
- 19:13learned a lot more quickly in this current moment.
- 19:16Yeah. Well, I'll just say, going back to the
- 19:18funding, is that I'd often say to funders, why
- 19:21do you think that disinformation only starts
- 19:2212 weeks before the election and then stops a
- 19:24day after elections? Why are you funding Brazil,
- 19:27then you fund Nigeria, then you fund France,
- 19:29then you fund Germany, then you fund the UK,
- 19:31as if none of this is connected, right? So I
- 19:34was very disappointed by that. And then your
- 19:36point, Elliot, is every time there was an election
- 19:37in the US, I tried to get funding to do projects
- 19:40in the US and never could because it was like,
- 19:42oh, we're so different. And of course they're
- 19:45not. But there is definitely a superiority complex.
- 19:47I say that as somebody who lives here and is
- 19:49now a citizen. But yes, it's frustrating that
- 19:51there's not more collaboration in the US. I think
- 19:56at this point, this would take us to about 2018,
- 19:59I think. And is this about the time you joined
- 20:02Brown? Is that correct? No, at that point, I
- 20:07was still leading First Draft full -time. We
- 20:10went to Harvard for a year to try and get funding.
- 20:12I mean, it was all this like moving around trying
- 20:14to figure out where we could be safe and what
- 20:17would work. But yes, it was a real challenge.
- 20:20And then really COVID happened and we got significant
- 20:24funding from Google, it will surprise nobody.
- 20:26A lot of money that was like, you have to spend
- 20:29that money in a year. So we kind of expanded
- 20:31and did all this work in 2020. And then that
- 20:34money went away in 2021 and we had to lay off
- 20:37lots of staff. And at that point, you know, attacks
- 20:41against those of us that were doing this work
- 20:43were increasing. We couldn't get publishing insurance
- 20:45because people like we can't like we can't support
- 20:47a site that publishes misinformation because
- 20:49you're going to get sued. And so it just became
- 20:51increasingly difficult for us to do that work.
- 20:53And so at that point, I took some staff to Brown
- 20:56University, but it was really kind of the end
- 20:58of first draft. And I think, Charlie, you put
- 21:01the link in. And then lovely Internet Archive,
- 21:06who of course we love and we all use the Wayback
- 21:07Machine and we have their things. I contacted
- 21:10them and said, look, there's such important work
- 21:12on here. Is there any way you could help? And
- 21:14they were like, oh yeah, we'll archive in perpetuity.
- 21:18And so I had this wonderful couple of weeks where
- 21:21I had to go through all of the archives of First
- 21:23Draft and then pick out how to reorganize the
- 21:26site to make it a place that people could go
- 21:29and find information. But like I say, it does
- 21:32sit to me like an archive that somebody one day
- 21:34will go through and study because there's an
- 21:36extraordinary depth and wealth of materials,
- 21:38both articles and blog posts, but also training
- 21:40materials that were translated into many, many
- 21:43languages. So yeah, I will always feel very warm
- 21:47and fuzzy about First Draft because not only
- 21:48was it started by a really great group of people,
- 21:50we ended up I think doing work that really helped.
- 21:54I think that period around COVID is one where
- 21:58now with the rise of this censorship industrial
- 22:01complex theory has become a real point of contention.
- 22:05I think that was a moment where traditional institutions
- 22:09were faced with a huge crisis. They struggled
- 22:13to adapt because they were still thinking of
- 22:15it in terms of their old institutional dominance
- 22:17of the information environment. They were relying
- 22:20a lot on academics and independent organizations
- 22:23to fill in the gaps that they weren't able to
- 22:26handle. I think as well post -2016, there was
- 22:29so much focus on the idea of bad information
- 22:31versus good information. and that the kind of
- 22:35foreign interference had become such a big part
- 22:37of the discourse, that really the more underlying
- 22:40factors were missing there. I'd be here to hear
- 22:43your kind of experiences and reflections in that
- 22:45specific period, because that then came back,
- 22:48I think, to haunt a lot of us, you know, in the
- 22:50new Trump administration. Yeah, no, I have a
- 22:54lot to say about this. mean, I'm so glad we're
- 22:56having this conversation. But I also have an
- 22:59eye to these conversations aren't happening in
- 23:01public because there's this sense of if we were
- 23:04to kind of acknowledge some of the missteps,
- 23:06it would look like that we were kind of giving
- 23:08oxygen to certain ideas. But I look back at COVID.
- 23:11I mean, even the first draft team, the majority
- 23:13of people in their 20s or early 30s, the majority
- 23:15of people were living alone. hadn't had a cuddle,
- 23:19myself included. I look back now and we were
- 23:21all working full time monitoring conspiracy theories
- 23:23and everybody was trying to do good. Did that
- 23:27mean that we labelled stuff as misinformation
- 23:29that was a poor use of that term? 100%. There
- 23:32was stuff that we labelled as misinformation
- 23:34when we should have said this is unsettled science,
- 23:37or we don't have enough evidence yet, or we're
- 23:39concerned that people are pushing this narrative
- 23:41to make money. you know, we need more scientific.
- 23:43We didn't do that, right? People were looking
- 23:46for, is this misinformation or not? We were trying
- 23:48to ensure that the platforms who had created
- 23:51these policies, were they actually doing what
- 23:54they said they were going to do in the policies?
- 23:56So to your note about the censorship industrial
- 23:58complex, like... there's a blog post that kind
- 24:01of lists the 50 organizations in the censorship
- 24:03industrial complex and like first draft was number
- 24:05one, right? And I remember reading that, I mean
- 24:07like, but what that was, was just like a number
- 24:10of people trying to figure out how to do this
- 24:12work. But the platforms, they didn't really go
- 24:14after the platforms, right? We weren't calling
- 24:16up Google or Twitter, you need to take this down
- 24:19because we think it's bad information. We're
- 24:21like, hey, your policy says X and here's an example
- 24:23of something and we're just letting you know,
- 24:25it's up to you to do whatever you want, but this
- 24:27goes against your policies. And that's got lost
- 24:29in all of this, right? Like it was, it's about
- 24:31these individual evil academics or nonprofits
- 24:34as opposed to actually these were the platforms
- 24:36that were making these kinds of decisions. But
- 24:38I do look back and I think, I wish we hadn't
- 24:41been moving at speed. I wish we'd had more time
- 24:43to collaborate, but we couldn't because we were
- 24:44all living alone or not able to connect. And,
- 24:47you know, I think that we didn't think through
- 24:49that this would look like, we didn't think about
- 24:53the balance of freedom of expression. like which
- 24:56we all believe in, and how do we ensure freedom
- 24:59of expression exists when we have speech that
- 25:01is potentially harmful, right? And we still haven't
- 25:03figured out that balance, right? Policymakers,
- 25:06you know, international human rights lawyers
- 25:07haven't figured that out either. And so I think
- 25:10organizations like First Draft and other nonprofits
- 25:12in this space like Center for Countering Digital
- 25:14Hate that are getting absolutely hammered, like
- 25:16all of us were trying to do the right thing.
- 25:18Was it perfect? Were mistakes, of course, but
- 25:21there were mistakes made by all sorts of people
- 25:22in all sorts of ways. But we were really good
- 25:24people trying to do the right thing, right? But
- 25:25we've been painted as these like evil monsters.
- 25:28And that's pretty hard to swallow. I think a
- 25:33big problem with this as well is the failure
- 25:35to recognize the change in the information system
- 25:37and how freedom of speech is built around information
- 25:41systems that were dominated by institutions,
- 25:44by governments, by media, that pre -filtered
- 25:47certain views and ideas out of discourse. The
- 25:50guy standing on the corner of the street, holding
- 25:52a sign saying the earth is flat, didn't get onto
- 25:54the news at one BBC. You live in a different
- 25:57information environment now. In my work on verification,
- 26:01depression and accountability, I like to make
- 26:04the point that those processes that were done
- 26:06by institutions have kind of been collapsed into
- 26:09the public space now. The public aren't just
- 26:11recipients of information, we're distributors
- 26:13of information, yet we don't hold the responsibility
- 26:16of making sure that information is accurate.
- 26:18We can just share what the hell we want. I think
- 26:20that's where that tension starts being created
- 26:23because if you start saying, oh, well, people
- 26:25can't say this because they can't say that because
- 26:28it's wrong and it's untruthful. Even if you can
- 26:29prove that's wrong and untruthful, it's still
- 26:32you limiting their speech. But as an academic,
- 26:36as someone working at Ballincat, we aren't the
- 26:38government, which is really a lot of what freedom
- 26:40of speech in the 20th century is focused on.
- 26:43Yet we have a problem now that governments are
- 26:45trying to adapt to that. So they kind of use
- 26:47academics and independent civil society organizations
- 26:50as a kind of proxy, and then use that to also
- 26:53pressure the social media companies into action.
- 26:55And I think there is a legitimate criticism of
- 26:58that. I think the censorship industrial complex
- 27:00reaches too far. And I think it's very biased
- 27:03in the way it's kind of approaching these topics.
- 27:04And I think it imagines stronger connections
- 27:07between our organizations that really, then it
- 27:10actually exists. It's not like we're on the phone
- 27:11all the time saying how we censor this because
- 27:13we don't like it. Yeah. And so for me, I think
- 27:17there is a legitimate criticism of that. I don't
- 27:20think the censorship industrial complex theory
- 27:22is that legitimate criticism. And it's now being
- 27:24weaponized against those people who certain groups
- 27:28found, you know, are basically part of the out
- 27:31group and the in groups now angry with them.
- 27:33I think in this current administration, that's
- 27:35become particularly dangerous. Yeah, and I mean,
- 27:38I think you get tired if you call out hypocrisy
- 27:40all the time. But exactly to your point, you
- 27:43know, when we think that's when people say to
- 27:45me, oh, yeah, but we just need to change Section
- 27:46230. It's like, can we think about what these
- 27:48different policies look like in the hands of
- 27:50different types of governments? right? Like this
- 27:52idea that there are like neutral policies without
- 27:54recognizing, well, who does make these decisions
- 27:56around speech? And so I think it's very frustrating
- 27:59to see, you know, those of us being attacked
- 28:01as being part of the censorship industrial, you
- 28:03know, I was being investigated by Jim Jordan
- 28:05and having my email sent to Congress, like seeing
- 28:08all of all of that about being deemed a censor,
- 28:11yet seeing like book bans in Florida, like, of
- 28:14course, it doesn't make any sense, right? That
- 28:15we're not thinking about the system as a whole.
- 28:17All of this has been weaponized for political
- 28:18gain. And so But I think those of us who got
- 28:22into this space were just internet nerds, trying
- 28:25to do the right thing. And then I have these
- 28:27moments of just like, how on earth did this happen?
- 28:29How did I go from working with the BBC about
- 28:32how to verify what Beryl from Swansea is sending
- 28:35in to being investigated by Jim Jordan? But that's
- 28:38this insane journey we've been on. But we have
- 28:41to understand it, to your point, Elliot, from
- 28:42this very much higher level zoom out of like,
- 28:46it's not about the debunking individual claims
- 28:48or what was labeled or what wasn't. It's like,
- 28:49what's happening at the institutional and ecosystem
- 28:52level that really explains all of these changes?
- 28:55And I think as we're moving into this current
- 28:57period, I mean, as you've mentioned, you've been
- 28:59targeted. I think a lot of our colleagues, you
- 29:02know, people we've worked with in the past, people
- 29:04who are just part of the kind of counter disinformation
- 29:06community are finding an incredible amount of
- 29:08pressure. I mean, we all know about the MIT technology
- 29:10reviewer report where emails have been sent out
- 29:13asking for communications from certain individuals.
- 29:16You know, Ename and Balincat was also mentioned
- 29:18within that. And that in itself is a potentially...
- 29:22you know, it pushes people out of the discourse.
- 29:24It makes it dangerous. How have you kind of felt,
- 29:28you know, in the last year with the kind of rising
- 29:32tide of this, your feelings of still being in
- 29:36the US, especially when things seem to be getting
- 29:38worse rather than getting better? Yeah, I mean,
- 29:41I'd be lying if I didn't think that I've had
- 29:43like wobbly moments. And I certainly before I
- 29:46was a citizen felt much more wobbly. And I mean,
- 29:50the reason that these things work is that then
- 29:51you have self -censorship, right? Like I was
- 29:53less likely to stand on stage and say certain
- 29:55things. I was less likely to write blog posts.
- 29:58I think now I've got to the point of like, I
- 30:00don't want to hide anymore. And, you know, and
- 30:02just, but I mean, I think, and it's hard to,
- 30:05I mean, I'm sure you get horrific emails. I mean,
- 30:06I get horrific emails, which say horrific things
- 30:09to me. And there were times I was like, I just
- 30:11don't want to deal with it. But I think I've
- 30:14come through another side now and just like,
- 30:17sod it. I've got a lot to add. And I think it's,
- 30:19but I also just want to acknowledge that. Um,
- 30:21women in particular get harassed in pretty nasty
- 30:24ways. Um, but I do, uh, you know, I know it's
- 30:26trite to say, but it's too easy to let them win,
- 30:29um, by not putting a head above the parapet and
- 30:32saying the things that you need to say. And I
- 30:34think it's difficult to have conversations about
- 30:36this in public because on one side, you have
- 30:38people are really just trying to push a specific
- 30:40agenda that, you know, we're the sensors and
- 30:43the censorship industrial complexes is evil.
- 30:45And if you try and kind of meet them halfway,
- 30:47they aren't interested in that because they're
- 30:48absolutists in their positions. And I think there
- 30:50should be a debate and discussion about where
- 30:53things went wrong. But if it just results in
- 30:56us being attacked more, then it's very hard to
- 30:58have those discussions in public. And Charlotte,
- 31:00how are we doing for questions from the audience?
- 31:04A couple of people have asked about AI, actually.
- 31:09Rope Stone wrote, if governments are weaponizing
- 31:11disinformation to the public, How can the consumer
- 31:14in the street stand a chance of finding the truth?
- 31:17Will the trend of LLM and AI give way to much
- 31:20information for the user to comprehend what is
- 31:22actual truth? And as a secondary to that, Sarah
- 31:25said, is there any old school verification techniques
- 31:27that can sidestep this new trend in people trying
- 31:33to use AI to verify things? Yeah, I just taught
- 31:36a class just now actually on generative AI, good,
- 31:39bad and ugly. And I was trying to say to the
- 31:41students like the thing that's the ugliest part
- 31:43for me is the fact that people are losing trust
- 31:45in anything they see or hear. Right. And we saw
- 31:48that the first time I saw that was around Gaza
- 31:50when I saw people like that's so upsetting that
- 31:52can't possibly be true. But seeing yesterday
- 31:55with the Alex Pretty video coming out of like
- 31:59you know, him potentially hitting a car and spitting,
- 32:01and seeing people like, that can't be true, that
- 32:03can't be true. And I was like, yeah, that's the
- 32:05biggest victim of this moment. Um, but I think
- 32:08what's so interesting about AI verification is,
- 32:10of course, there's all these startups trying
- 32:12to convince us all that they've got the technology
- 32:13that can do this automatically. But in the same
- 32:16way as we started when we were doing this work,
- 32:18you know, 15 years ago, there was never one easy
- 32:21tool, right? You had to, like, look at the exif
- 32:23data and then do the shadow analysis and then
- 32:25check the timestamp on the, like, there were
- 32:27all these different checks you had to make. And
- 32:28that's what I say with AI, which is like, the
- 32:32stuff that you see, you've still got to do the
- 32:33human checks of like, did this, do you have evidence
- 32:35that this happened independently? And if you
- 32:38don't keep quiet, like there are times that newsrooms
- 32:40were like, we've got 99 % there and the newsrooms
- 32:42are like, but we're not 100 % there. So we're
- 32:44not going to use that footage. But I think we're
- 32:46so far away from that now that that's my concern
- 32:49about AI. But I do hope that with this new generation
- 32:52of students and people who have grown up with
- 32:55AI, we also need to keep saying if we all lose
- 32:58trust in everything, we're absolutely screwed.
- 33:00And B, here's some fun techniques. So when people
- 33:03say, well, what's the magic AI verification tool?
- 33:05I'm like, here are the magic non -AI verification
- 33:09tools that still stand the test of time, which
- 33:11is like picking up the phone and talking to somebody
- 33:13and be like, were you there? Were you an eyewitness?
- 33:15Can you tell me what you saw? Sometimes there's
- 33:18just no shortcut. So yeah, I'm interested to
- 33:20hear what Elliot says. And I think on that example
- 33:23you gave of the recent video of Alex Petty kicking
- 33:26the... A lot of people I saw reacting, oh, that's
- 33:30AI, that's not real, it's been manipulated. That's
- 33:32no different when Trump was showing the pictures
- 33:34of the video of the things being thrown out the
- 33:36White House window and he said, oh, that's probably
- 33:38AI, it's not real, it's AI. It's the same kind
- 33:41of mindset that it's about protecting the group
- 33:43identity rather than verifying something. And
- 33:46if he kicked that car, it doesn't justify him
- 33:48being shot to death two weeks later. Not all
- 33:52the people using it as justification are either
- 33:54bootlegs or psychopaths, I think. So But it is
- 33:57about, you know, group identity for a lot of
- 34:00people on both sides of these situations. And
- 34:02I think we need to commit ourselves primarily
- 34:04to, you know, building our group identity around
- 34:07that verification, deliberation, accountability
- 34:10process being functional rather than serving
- 34:12the interests of the group. Yeah. And I do think
- 34:15whenever, you know, I talk to students and others,
- 34:17there's like, here's the technical verification
- 34:19stuff, but here's why our brains are mush. like
- 34:22left or right, like this is why we're vulnerable
- 34:25to this. And the in -group, out -group, the fact
- 34:27that we're all performing when we're on social
- 34:29media, that's the stuff that I think doesn't
- 34:31get taught as much when we do a checklist of
- 34:32like, what's the pope wearing a puffer jacket?
- 34:34But yeah, all right, we can do those checks.
- 34:36But actually, why might you be convinced? Well,
- 34:39because you looked great in that puffer jacket.
- 34:40And you want it to be true because it's funny,
- 34:41right? Like that's a silly example. But we don't
- 34:43talk enough about the psychology of all of this.
- 34:50platforms when you were talking and a couple
- 34:53of questions have come in that. When did the
- 34:55tensions between journalists, researchers and
- 34:58platforms really emerge when it comes to disinformation?
- 35:03And Nick asked, if governments, can governments
- 35:05regulate platforms to make algorithmic curation
- 35:08less alien in the attention economy? So the first
- 35:14question, I mean, I have to say as somebody who's
- 35:16probably been to like 400, 500 counter disinformation
- 35:20conferences, I mean, that's where the industrial
- 35:21complex was just like event after event after
- 35:23event. And people would always invite the platforms
- 35:26and were like, yeah, we got Google to come, or
- 35:28we got Facebook to come. And the representatives
- 35:30would turn up with their marketing slides that
- 35:32had been signed off by the... And they said nothing,
- 35:35right? But we were like, oh, we should keep inviting
- 35:36them. I think the shift happened in 2021 when
- 35:40everybody realized post -COVID that there was
- 35:43a shift happening and there was fears around...
- 35:46There was starting to be these grumblings that
- 35:48there had been jaw boning, the idea that the
- 35:50Biden administration had been... putting pressure
- 35:52on the platforms. And it was around that time
- 35:55that people really stopped talking to platforms
- 35:58because there was fears of like collusion and
- 36:00jawboning. But that has really, I mean, I used
- 36:02to speak to almost everybody, like lots of people
- 36:04at all platforms a lot of the time and like all
- 36:07of those relationships have disappeared. And
- 36:09then of course, Musk took over, trust and safety
- 36:11teams were demolished. But yeah, again, it's
- 36:14kind of very interesting to track like Google
- 36:16funding a first draft. And the other day, on
- 36:19my reading list, I had a piece of research from
- 36:22Jigsaw about toxicity online. And my student
- 36:25was like, hey, Claire, I can't find the PDF.
- 36:26And I was like, oh, here's a link. And I was
- 36:27like, oh, it's been taken down. I mean, Google
- 36:29is so frightened that some of their best research
- 36:32has been taken down off the internet because
- 36:33they don't want to upset Trump, right? So it's
- 36:35an extraordinary journey that we've been on.
- 36:38The question about regulation is a really, really
- 36:40good one. I used to believe that we would get
- 36:43someone with regulation. And I was very excited
- 36:44about the Digital Services Act from the EU. that
- 36:47there was a real emphasis on data transparency,
- 36:50a real emphasis on auditing algorithms as opposed
- 36:53to content. high hopes. And then of course, as
- 36:56we've seen, since Trump has come back into office,
- 36:59pressure from the US against the EU on these
- 37:03fines. And I worry that there will be some element
- 37:06of pulling back and a chilling effect when I
- 37:08think if Trump hadn't got in, we would have seen
- 37:10like some pretty brave moves from the EU against
- 37:14the platforms. I do not see that happening. I
- 37:16wish I could be wrong, but I'm not as hopeful
- 37:20as I was two years ago about what regulation
- 37:22might be able to do. And I would add to that
- 37:26that we've had the same experience. I mean, up
- 37:29until Musk took over Twitter, I mean, like literally
- 37:31as he took over, we were just starting a new
- 37:34project with Twitter about looking kind of identifying
- 37:38disinformation that works and stuff like that.
- 37:41And literally the team's emails stopped working
- 37:43overnight once Musk took over. And I had a lot
- 37:46of people I knew very well at Twitter who had
- 37:48been really working on these issues seriously.
- 37:50just kind of disappearing. Like, you know, as
- 37:53we all contact each other by email and then their
- 37:54emails work, it's like, I don't have their phone
- 37:56number or any of the details for them. So yeah,
- 37:58and it was, it felt quite sudden when it happened,
- 38:00not just in terms of X, but just that, like you
- 38:02said, the post 2021 period. We were talking a
- 38:06second ago about the U .S. Sarah said, are you
- 38:09seeing any changes in U .S. organizations going
- 38:12from, no, we don't need any help researching
- 38:14disinformation, to we really need help researching
- 38:17disinformation? Or are you seeing the opposite
- 38:20because of Trump in power? Yeah, I mean, the
- 38:25sad thing is, is that kind of three years ago,
- 38:27there was a really thriving network of research
- 38:30institutes and nonprofits in the US that were
- 38:32being funded and supported. And that network
- 38:35has almost entirely gone right from the project
- 38:40at Stanford with Ray DeResta and Alex Deimos.
- 38:45Stanford basically once they were attacked. shutting
- 38:48it down, you know, Joan Donovan at Harvard, we
- 38:50were doing work at Brown, you know, like it's
- 38:52first draft has gone, like it, there's many,
- 38:54many, many organizations that could just couldn't
- 38:56get any funding anymore, or were basically attacked
- 39:00into, and they're bigger organizations saying
- 39:02when we, you know, we can't support you anymore.
- 39:04So it is extraordinary to me to see the collapse
- 39:07of a pretty tight knit. infrastructure of counter
- 39:11disinformation efforts in the US completely disappear
- 39:14at a time when we need them more. But that was
- 39:16the whole purpose. That was the whole purpose
- 39:18and it was successful and it stopped people from
- 39:20doing the monitoring. You know, we're going to
- 39:21go into the midterms. In fact, in the 2024 elections,
- 39:25we used to have a whole network of people that
- 39:27would do these big election projects and would
- 39:29be monitoring. And all of that has disappeared
- 39:31now. So we just don't have the monitoring infrastructure
- 39:33that we used to have. So yeah, it's pretty shocking.
- 39:37I'd add that as well. The cut to USAID funding
- 39:41has had a devastating effect outside of the US
- 39:43as well. We know loads of media organizations
- 39:48who are using USAID funding, a lot were very
- 39:50reliant on it. They've all started to collapse
- 39:52and disappear. But also that has an impact on
- 39:54the wider funding ecosystems. It means there's
- 39:57less money in general for organizations who are
- 40:00doing this kind of work and priorities are changing
- 40:02a bit. And I think also kind of funders are starting
- 40:07to feel that maybe this counter disinformation
- 40:09stuff, these fact -checking websites aren't actually
- 40:11as impactful as they think they are. And I think
- 40:14they are correct to an extent on that. There's
- 40:17no money now to figure out the more effective
- 40:19version of kind of responding to these issues.
- 40:22And, you know, that's affecting every organization.
- 40:24It certainly affects Bellingcat, even though
- 40:26we never took USAID funding, we're still competing
- 40:28in that much smaller pool of money. And there's
- 40:32also been a huge number of journalists, researchers
- 40:34made redundant from these roles. So if you're
- 40:37working in it, it's very difficult to find any
- 40:40kind of employment. And you are seeing, I think,
- 40:42a bit of a rise in more independent outlets,
- 40:45but still they've got the challenge of having
- 40:46a subscriber base or, you know, hoping Substack
- 40:49works with them or competing in this dwindling
- 40:51pool of money. I was just going to talk about
- 40:58platform money, right? First off, we definitely
- 41:00got criticized for taking platform money. My
- 41:02perspective was always, we got money to do good
- 41:04work, right? And there was always editorial independence.
- 41:07And why shouldn't they pay to help clean up the
- 41:10mess that they'd created? But even seeing how
- 41:12Meta has pulled back, I mean, collapsing fact
- 41:15checking in the US. And we can have a longer
- 41:17conversation about the efficacy of fact checking.
- 41:19But fact checking when you're doing it on Meta's
- 41:21platforms and the reach that those fact checks
- 41:23got once they retrained the algorithms to basically
- 41:26to lower certain types of content or to flag
- 41:28others. But yeah, there are many, many fact checkers
- 41:31that started because of money from meta that
- 41:34have now, in many, many parts of the world, that
- 41:36we no longer have fact checkers. Have we in the
- 41:41past overestimated the power of kind of corrections
- 41:44and fact -checking? Kouch von Katschi asked a
- 41:48question that kind of fits in here. How do we
- 41:50go about teaching how to navigate this new information
- 41:52landscape that is very fractured and has decreasing
- 41:54trust? It feels like we're in a massive information
- 41:57and geopolitical paradigm shift, and I'm curious
- 41:59how we work through this. So if immediate fact
- 42:03-checking isn't the way forward now, what would
- 42:06you suggest is? So I'll be transparent in that
- 42:11I'm a trustee of Full Fact, which is a UK fact
- 42:13-checking organisation. I've always been clear
- 42:16that fact -checking in terms of changing an individual's
- 42:20belief system is very, very difficult. Fact -checking
- 42:22is incredibly important and even more so now
- 42:24that we have an accurate record. In an age of
- 42:26LLMs, it's even more important that we actually
- 42:28do have an archival moment that says what something
- 42:31was, right? So for me, it's about having an accurate
- 42:36record and it's also at times ensuring that the
- 42:39person who lied in the first place changes it.
- 42:41They're very good at getting people to correct
- 42:42the record. it is all really important. But to
- 42:46Elliot's point, we spent so much money on the
- 42:49end of the information supply chain, right? Like,
- 42:51let's wait until people are seeing the bad stuff
- 42:53and then we'll kind of issue a correction. We
- 42:54should have been working much, much earlier in
- 42:56the supply chain. And we didn't do that. But
- 42:59that doesn't mean that fact checking doesn't
- 43:00work. It just doesn't work in the way that I
- 43:02think a lot of policymakers thought was like,
- 43:03oh, we just need people to have more facts and
- 43:05then they'll change their minds. But it is astonishing
- 43:08to me when you look at community notes on X and
- 43:10other places, how many times people linked to
- 43:12fact checkers. and like the LLMs, like if you
- 43:15go to ChatGBT, it pulls all of that stuff from
- 43:17those small fact checkers, has never paid a penny
- 43:19for that information, uses all of their archives
- 43:22to train their models. So for that now, I actually
- 43:24think the fact checking is more important than
- 43:26ever, but the funding is disappearing. I always
- 43:30find it amusing that Grok is a big fan of BedlingCat
- 43:33despite Elon Musk's position on the BedlingCat.
- 43:37I mean, for me, a lot of this is actually getting
- 43:41into that discourse as it's unfolding. And I
- 43:43think the recent shootings has been a really
- 43:45good example of that. We knew the US government
- 43:47was going to lie about the second shooting because
- 43:48they lied about the first one. So at Bellingham,
- 43:51we were very keen to make sure we could get information
- 43:53out very quickly. So we published stuff up through
- 43:55Blue Sky and on our other platforms without in,
- 43:58you know, hours of the shooting happening as
- 44:00that discourse was forming and we still were
- 44:02able to do our kind of longer form reporting
- 44:04later on. Also training other people to do that's
- 44:06really important. You had lots of media, traditional
- 44:10media doing open source investigation into that
- 44:13incident, partly because they've been trained
- 44:15to do that. Banning Cat has trained a lot of
- 44:17those people. We've created content online. You've
- 44:20done that kind of work. It's put those skills
- 44:22out there, but also involving the public. We
- 44:25asked a lot about this and say that this Discord
- 44:28community that we're talking to now played a
- 44:31role in getting that information out. Our trained
- 44:33volunteers community had a role in getting that
- 44:36information out. And by spreading those skills,
- 44:39it means there's more people applying them to
- 44:41this information, often in places that you aren't
- 44:43looking at. I always think of the role the kind
- 44:46of alternative health community played in spreading
- 44:49COVID disinformation. They weren't formed around
- 44:52COVID disinformation. They were formed about
- 44:54what were often traumatic experiences they had
- 44:56with medical professionals. And then they built
- 44:58a network around that, which towards more disordered
- 45:04forms of communication. We can't predict where
- 45:10this stuff can come from, so for me it's about
- 45:12spreading those skills as far and wide as possible
- 45:15through every layer of society so that when this
- 45:18stuff comes out, and as I mentioned before, we're
- 45:19all distributors of information, people at least
- 45:22think twice about the stuff they're sharing,
- 45:24and maybe that will stop some of the formation
- 45:26of these very strong opinions that are based
- 45:28on fictions. Yeah, and I'll just support that.
- 45:32Like I make jokes every time I speak about how
- 45:34I spend a lot of my days in the armpit of the
- 45:36internet. And so I spend a lot of time in spaces
- 45:38that are deeply participatory, that they're conspiracy
- 45:41communities or anti -vax communities. But they
- 45:43are communities where people turn up, they feel
- 45:45heard, they're told that they can do their own
- 45:47research, they have a sense of agency. Right.
- 45:49And the only community to me that matches those
- 45:52kind of communities is the Bellingcat community.
- 45:54I think if you think about trusted gatekeeping
- 45:57type organizations, they're still top down. put
- 46:00the right person behind the podium or you put
- 46:01out the right PDF, right? There's this idea that
- 46:04it's still top -down and hierarchical. And very,
- 46:07very few organizations, if any, on the kind of
- 46:10the side of quality information have said in
- 46:13order to get better information, you need collaborative
- 46:15spaces where people feel a sense of agency and
- 46:18you can work together, right? So just to say,
- 46:21I hope you will realize how important this space
- 46:23is, but it really is a model that I think many,
- 46:25many organizations, because they feel like they
- 46:26lose control if they hand it over. right? If
- 46:29they hand over messaging or communication, there's
- 46:32this idea that the quality goes down as opposed
- 46:33to, well, actually, if you hand it over, the
- 46:35quality will go up, but you're going to lose
- 46:37control. So I couldn't say more about how important
- 46:40this kind of community is. And I saw something
- 46:41in the chat being like, yeah, but these kind
- 46:43of communities are slower and other things. I
- 46:46do ultimately think we're in this really nasty
- 46:48moment right now where, like, how many times
- 46:51do you hear people go, oh yeah, well, I'm on
- 46:53Instagram. I mean, I know it's 75 % AI slop,
- 46:55but I still look at it, right? I think in a few
- 46:57years time, we'll actually have a kind of a bifurcated
- 47:00internet where it's not like I don't love AI
- 47:02slop sometimes. Like I want to see a cat singing
- 47:04and dancing, right? Because I'm in a bad mood
- 47:06and I want to be cheered up. But there are other
- 47:08times where I only want good information. And
- 47:09I think there will be increasing spaces where
- 47:12quality information coming from like quality
- 47:14organizations will have, you know, like watermarked
- 47:18content or they will have, you know, the kind
- 47:21of things that we're seeing around AI generated
- 47:23content that says this really was filmed by the
- 47:25New York Times or this really was filmed by or
- 47:27created by Bellingcat so that you can go to places
- 47:29for that kind of content. Doesn't mean we won't
- 47:32have the nasty sloppy parts of the internet,
- 47:34but I think people will realize that you can't
- 47:36have both in the same space. It shouldn't be
- 47:38on all of us individually to figure out is that
- 47:40true or not. So it could be wrong, but that's
- 47:42a kind of a prediction I have for the next five
- 47:44years. I'll just add to that. I had it interesting
- 47:47when I left X and then into Blue Sky. I wasn't
- 47:51planning to join Blue Sky, but to be honest,
- 47:53when I quit X, I felt like physically depressed.
- 47:56It was like I was so addicted. It's like going
- 47:57clawed into me so much. I thought I'd need like
- 48:00a methadone to get over this. And that's what
- 48:02Blue Sky became for me. But then when I actually
- 48:04started posting there, the discourse felt so
- 48:07different. There's so few times when someone
- 48:10would post something like banning Kat as the
- 48:12CIA or how terrible I was, which I got way too
- 48:15used to on X. And I think, you know, that having
- 48:20those good spaces where there can be discourse,
- 48:22where there is engagement, that you can talk
- 48:23to people normally, reminds me of what X used
- 48:26to be like, you know, or Twitter 15 years ago.
- 48:29Um, my worry is blue skies that, you know, as
- 48:32it grows, they're going to think, how can you
- 48:34squeeze some money out of this? And then it becomes
- 48:35an algorithmic hellscape. But I think we have
- 48:37to kind of, uh, you know, express the value of
- 48:41having those spaces where the algorithm doesn't
- 48:43dominate the conversation and that you don't
- 48:45get reshares just because you're sharing a, you
- 48:47know, singing and dancing cat video made with
- 48:49AI. Yeah. Well, I mean, I certainly think non
- 48:53-algorithmic spaces like WhatsApp or Reddit or
- 48:55Discord, you're seeing a growth of those kind
- 48:58of spaces because people are sort of fed up with
- 49:00the algorithmic stuff. There will always be a
- 49:04place, like I said, there's a role for that kind
- 49:07of content that makes me feel better when I've
- 49:08had a bad day. But there were also times where
- 49:10I just want good information in ways that I can
- 49:14verify who the person is. Yeah, Rafi agrees.
- 49:18They said, I think the overflow of slop will
- 49:20give people a renewed interest in controlling
- 49:22whether their feeds can be consumed with content
- 49:24that is not conducive to a useful flow of information.
- 49:27I just wanted to flag as well, above Kara said,
- 49:29just wanted to add fact checking can also be
- 49:31powerful if used by relatives, friends, act of
- 49:34those falling for disinformation. So it's also
- 49:36important, you know, to work within your local
- 49:39and your family communities as well. It's not
- 49:42just online. Yeah, but there's really good research
- 49:46about bystander fact -checking, right? So if
- 49:48somebody on Facebook says something bonkers and
- 49:51you respond, it's unlikely you're going to convince
- 49:53the original poster, but you have a lot of people
- 49:55lurking who see that and are likely to be like,
- 49:58oh, that's interesting. Let me go and read that
- 50:00article. So I often say to people, it's not about
- 50:02changing the mind of the first person, but recognize
- 50:05that you're having benefits in other ways. Absolutely.
- 50:09Nick has an interesting question earlier. Could
- 50:11governments see epistemic security as defense
- 50:14spending, like for example, education reform
- 50:16and journalism funding counting towards NATO
- 50:18spending targets? What do you think of that idea?
- 50:21Has there been any interest from governments
- 50:23in general to kind of pursue kind of fact checking
- 50:27or verification strategies? Is that being taken
- 50:29as seriously as it should be? Yeah, I mean, Elliot
- 50:33alluded to this earlier, which is working with
- 50:35governments is always really, really tricky.
- 50:36Like when we did our project in Brazil, we had
- 50:38to basically write a statement for a lot of newsrooms
- 50:41to say that we'd never taken money from the American
- 50:44government because of the American government's
- 50:45role in the military dictatorship. And so we
- 50:48never took money from the US government. We never
- 50:50took money from the UK government. And so I think
- 50:53there's this challenge of like, how can we stay
- 50:56independent? But your point is a good one. And
- 50:58there are like, for example, many of the Scandinavian
- 51:02countries and obviously the countries that the
- 51:04nearer you get to Russia, the more they take
- 51:05this seriously and see it as a kind of a counter
- 51:08disinformation as military spending. The question
- 51:11is like how much should they be doing versus
- 51:14independent organizations? And I think that's
- 51:16the real challenge with all of this. But I certainly
- 51:18think that there are many governments who are
- 51:21increasingly seeing this as something that they
- 51:23should be taking seriously. But I think they
- 51:24don't know where to start, right? So there's
- 51:25this sense of like, well, we should be doing
- 51:27more media literacy, but they don't know what
- 51:28format that should take or how to do it effectively.
- 51:31And it's very targeted on like school kids as
- 51:33opposed to like you need cradle to grave. So
- 51:36that's my frustration is that we haven't invested
- 51:38properly in innovative media literacy interventions.
- 51:42And instead we're like, well, we just would do
- 51:44a PowerPoint slide to some 18 year old or, you
- 51:46know, 16 year olds who laugh at the teacher who
- 51:48hasn't got a clue what Snapchat is. But yeah,
- 51:51I'm sure Elliot's got some thoughts on all of
- 51:52this. Yeah. And on that point, even regards to
- 51:54education, there was a recent government advice
- 51:57to curriculum development organisations about
- 52:00having more media literacy skills, which really
- 52:03for me is critical thinking skills needing to
- 52:05be embedded across multiple subjects. You're
- 52:08no longer in a society where if you want to influence
- 52:10discourse, you have to, you know, have a career
- 52:12path to become a journalist or politician. You
- 52:14pick up your phone and you're influencing the
- 52:16discourse just by looking at a post with five
- 52:18seconds longer than another post. We need to
- 52:21understand that's the environment we live in.
- 52:22We need to educate around that environment. And
- 52:24if we fail to do that, then we're just going
- 52:25to keep on repeating the same mistakes. It's
- 52:27like, um, you know, someone says, Oh, I go into,
- 52:30you know, do media literacy skills and we teach
- 52:32kids how to fact check headlines, which is things
- 52:34that have been said to me. I'm thinking, what
- 52:36headline are they seeing to ever fact check?
- 52:38Because that's not how they're consuming the
- 52:40content they're encountering. You don't have
- 52:42this kind of passive consumption of media in
- 52:45the kind of 20th century way where you had the
- 52:47newspaper that your parents brought home and
- 52:48the Scott News was on. now connected to a media
- 52:53ecosystem that's algorithmically designed to
- 52:55keep you on those platforms. We need to educate
- 52:57around that. We can't just say, oh, here's how
- 52:59you fact check a bunch of stuff because we need
- 53:01to go beyond that. We need to say, actually,
- 53:03democracy relies on these things actually working.
- 53:06We also need to be able to demonstrate that institutions
- 53:09respond to that. We're teaching young people
- 53:11that this is how democracy works. This is why
- 53:14facts are important and then the government comes
- 53:16along and does Iraq in 2003 all over again. That
- 53:20just drives the cynicism that sends people to
- 53:23these online communities that say, yes, you're
- 53:25right to hate the government for being liars.
- 53:27Here's lots of information to support that. And
- 53:29then in another 20 years where we have enough
- 53:31chemical weapons attack in the Middle Eastern
- 53:33country, we'll have again people saying, oh,
- 53:35well, it's just like it was 20 years ago. I think
- 53:37that's the responsibility here. We talk a lot
- 53:42about how people have lost trust in institutions,
- 53:45but that's often earned. It isn't just because,
- 53:48oh, it's a vague idea that they're reading too
- 53:50much rushing propaganda or bad information on
- 53:52the internet. It's because those institutions
- 53:54have failed people in a myriad of ways and those
- 53:57institutions need to be able to respond to that.
- 53:59But I think they also end up kind of paralysing
- 54:01this current information environment because
- 54:03What they see is the most salient discourse on
- 54:06these platforms that they're engaging with. In
- 54:08the UK it's, well, in the US as well, it's immigration
- 54:10because it's salient because one side is really
- 54:13passionate and hateful, as is the other side,
- 54:16against that side. So there's lots of arguing
- 54:18online. The algorithm sees, hey, great, loads
- 54:19of engagement and starts boosting that. That
- 54:22changes the incentives of people on the platforms
- 54:24to start posting more stuff about that topic.
- 54:26And they have the institutions who exist in those
- 54:28same spaces now seeing that as the topic of conversation.
- 54:31And they have people like, know, labour chasing
- 54:34after reform voters and actually pissing off
- 54:36all their own supporters. So we need to understand
- 54:39the information environment we're existing in
- 54:41even before we start discussing those solutions
- 54:44in a really deep way. I'm glad to say that I'm
- 54:46consulting with a couple of education establishments
- 54:49that are developing curriculum material and they're
- 54:51really interested in this kind of deeper way
- 54:52of approaching it. So I think we're going to
- 54:55get away from the kind of fact checking headlines
- 54:57in one lesson that I approach, which I think
- 55:00is something that's really, really important
- 55:01in this moment. And I think just, I saw somebody
- 55:05put into the chat about that blog post from LSE
- 55:07and it was from 2018. I remember in lots and
- 55:10lots of conferences, 2016, 2017, anybody, somebody
- 55:13talked about media literacy, somebody else were
- 55:15like, oh, we haven't got the time. We just need
- 55:16to get Facebook to tweak its algorithm or we
- 55:18just need a spam folder for misinformation. We
- 55:20haven't got time. And now we're 10 years on,
- 55:23I'm like, what would have happened if we'd actually
- 55:24invested in media literacy properly 10 years
- 55:27ago, right? Not to say that we wouldn't be where
- 55:29we are today, but it wouldn't be quite as bad
- 55:31as it is. And I'm with Elliot that we just haven't
- 55:33done enough to explain how algorithms shape what
- 55:36we see. Like I'm doing a project right now on
- 55:37polarization. And when you talk to people like,
- 55:40well, you know that we have like polarized candidates
- 55:42like Marjorie Taylor Greene or, you know, left
- 55:45-wing candidates because the algorithm privileges
- 55:47the more extreme you are. Like if you're a nice,
- 55:50centrist candidate, you don't get any airtime.
- 55:52So that's forced people to become more and more,
- 55:55you know. So once you start explaining how algorithms
- 55:57have impacted our headlines, the types of politicians
- 56:00we see, what politicians say, then I think people
- 56:03start to be like, oh, okay, right. So it's algorithms
- 56:06are at the heart of so many different things,
- 56:07but we've done a really bad job of explaining
- 56:09how algorithms work. And most people are like,
- 56:11oh, everything I see on Facebook is just the
- 56:13people I like. They still don't understand that
- 56:15what they see on Facebook is algorithmically
- 56:16driven and that they don't see 90 % of the stuff
- 56:19that they could. Then people's minds are like,
- 56:20ooh, and it's 2026. So yeah, algorithms. I'll
- 56:29put all of the links that have been mentioned
- 56:30in the chat in the description of the podcast.
- 56:32So if you are listening back, you can still check
- 56:34them out, including a recent article that Elliot
- 56:38wrote for the Voters Institute, which talks about
- 56:41how to protect democracies from falsehoods and
- 56:44has some examples of how you could teach critical
- 56:47thinking. So please do give that a read. Bear
- 56:52in mind, we only have about seven minutes left,
- 56:54so we have to limit the questions. I wanted to
- 56:58go back to the idea of kind of how we talk about
- 57:03disinformation. And Caro asked earlier, I remember
- 57:07reading Claire that you've become rather critical
- 57:09of the term malinformation. Can you elaborate
- 57:11on that a little? Are the terms that we're using
- 57:14disinformation, misinformation, malinformation,
- 57:16fake news, are they confusing, misleading? Are
- 57:19they causing more problems than they should?
- 57:24Yes, I wrote a piece in the spring of 2023, where
- 57:27I basically said that So I wrote a report in
- 57:302017 called Information Disorder, which came
- 57:32from a frustration that everybody was using the
- 57:35term fake news, which had been weaponized and
- 57:37actually didn't describe a lot of the problems.
- 57:39A lot of this content isn't fake and it certainly
- 57:41doesn't look like news. So I wrote this report
- 57:43and it had like it was a very long report, but
- 57:45it had this Venn diagram in around misinformation,
- 57:48disinformation, malinformation. And that Venn
- 57:50diagram has appeared on a lot of slide decks
- 57:52over the years. Like I've seen them pop up and
- 57:54I'm like, oh, that slide deck. But I sort of,
- 57:56I wrote this piece to say that wasn't helpful,
- 57:59right? Because we got obsessed with trying to
- 58:01define different examples. Like, but what bucket
- 58:03do we put it in? Right. And so much of this was
- 58:05about intent and we couldn't actually tell. And
- 58:08so this piece that I wrote was basically saying
- 58:09we, by doing that, we got fixated on thinking
- 58:12that we could take the complexity of the ecosystem
- 58:15and make it simple. But, and it meant that we
- 58:17completely forgot to look at a whole host of
- 58:19different factors, right? And so I talk about
- 58:22missing disinformation, right? Like, so if I'm
- 58:24in an anti -vax community and there's an image
- 58:26I have from Instagram with a woman who's holding
- 58:29a binder full of scientific evidence, right?
- 58:31She's reading the evidence. She's not somebody
- 58:33who just like forwards something in the coffee
- 58:35shop without checking. She's looking at this
- 58:37stuff all day long, but she's just reached different
- 58:39conclusions to me. Is she a disinformer or is
- 58:42she a misinformer? She's neither of those things,
- 58:44right? The people who stormed the Capitol. So
- 58:47the more I was doing this work, I was like, none
- 58:49of those early definitions make sense when you
- 58:51start operationalizing it. Or the people that
- 58:53use the, when Musk changed the blue tick thing
- 58:55to say it was Eli Lilly and said, now everybody
- 58:58can get like free. insulin. What was that? Was
- 59:03that misinformation or disinformation? Was it
- 59:05causing harm? Well, it caused harm to some very
- 59:06rich fat cats on Wall Street, but it was helping
- 59:09the public interest. So I think that these definitions,
- 59:13maybe they were necessary in 2016 to take a complex
- 59:16problem and make it make sense. But I actually
- 59:18really try and avoid a lot of the terms now,
- 59:20partly because they've been weaponized. But secondly,
- 59:23I don't think any of this is as straightforward
- 59:25as we like to think. And I think by doing so,
- 59:29it has prevented us from really understanding
- 59:31some of the key drivers to these problems. And
- 59:33so for that reason, I've changed my language.
- 59:36So yeah. And I do use those terms in my discourse
- 59:40framework, but that's because I can then build.
- 59:43So I use those as the elements and then I explain
- 59:46what the drivers are, who the people are, you
- 59:48know, the types of people. groups or people who
- 59:50are spreading this stuff and the process behind
- 59:52that. So that's, I think, the thing that's often
- 59:55overlooked. It has to be about when we're talking
- 59:56about this information, about the discourse,
- 59:59not just about, you know, this is bad information
- 1:00:02and it's getting out to people and therefore
- 1:00:03we have to make the information. It's like...
- 1:00:06Why does this discourse emerge in the first place?
- 1:00:09Because that lets us off far too lightly and
- 1:00:11it lets off institutions far too lightly. Because
- 1:00:14again, it comes back to this idea old people
- 1:00:16have lost trust in institutions, but we need
- 1:00:18to understand why they've lost trust in institutions.
- 1:00:21What was that trust even based on in the first
- 1:00:23place? We don't really ever question that. Um,
- 1:00:25and if we don't do that, then we'll never actually
- 1:00:28be able to find workable solutions. And I think
- 1:00:30that's where, you know, a lot of post 2016 counter
- 1:00:33disinformation efforts did fail because it was
- 1:00:35so much about the information rather than the
- 1:00:38discourse that was happening and the role of
- 1:00:39institutions in being responsible for that discourse
- 1:00:42in the first place. Yeah. And I think just to
- 1:00:46jump in, the question actually was about malinformation,
- 1:00:49which was this term that I coined with my co
- 1:00:51-author, Hossein Darikshan. And the idea was
- 1:00:54when look, genuine information is used to cause
- 1:00:57harm, right? And so I've done work with the ICRC
- 1:01:00and the UN, like some very effective malinformation
- 1:01:03is when you go back and talk about the ICRC's
- 1:01:06role during World War Two, right? Like, or you
- 1:01:08talk about the UN and what some of the peacekeepers
- 1:01:10have done, right? Like that is not false information.
- 1:01:13It's like that is it's being is the weaponization
- 1:01:15part that is useful and interesting. labeling
- 1:01:18it isn't that helpful, right? And so I think,
- 1:01:20but what I was trying to make the point by talking
- 1:01:22about malinformation is that everybody was so
- 1:01:23obsessed with false information, you could only
- 1:01:25flag false information without realizing that
- 1:01:27the smartest actors were weaponizing genuine
- 1:01:30information because you couldn't go to a platform
- 1:01:33and be like, you want to take that down, it's
- 1:01:34disinformation because it's not, it's genuine
- 1:01:36information, right? And so again, to Elliot's
- 1:01:39point, so much of this is understanding historical
- 1:01:40context, understanding institutions properly
- 1:01:44and the discourse. around them to understand
- 1:01:46how they're being weaponized. It's not about
- 1:01:47is it true or false. And that's, I think, where
- 1:01:49we got lost a little bit. Thanks for giving that
- 1:01:53extra context. We're going to have to wrap in
- 1:01:55a second, but I wanted to just read out this
- 1:01:56point. One of somebody in the chat said, 30 years
- 1:01:59ago, my views aligned with today's Marga. Many
- 1:02:02of my friends from those days are Maganal. It
- 1:02:04was people sharing facts as seeds, not demands
- 1:02:07that helped me change, that led me to be the
- 1:02:09centre -left pro -facts person I am now. So really
- 1:02:13builds on a lot of what you were saying. Just
- 1:02:17to wrap up, a lot of people have been asking
- 1:02:19how they can get in contact with you both if
- 1:02:22they have follow up questions or want to highlight
- 1:02:26amazing things that they're working on to you.
- 1:02:30Where's the best way to contact you? I'm going
- 1:02:34to put my email in the chat. So I might not be
- 1:02:38able to get back immediately, but I always love
- 1:02:39to hear people and what they're doing. And so
- 1:02:41it would be my pleasure to get into a conversation
- 1:02:44with anybody if they want to continue it. Amazing.
- 1:02:47You can contact me directly at my email as well.
- 1:02:50Fabulous. Thank you so much both for your time
- 1:02:53today. And thank you to everybody who's joined
- 1:02:56the audience as well. Thank you for your fantastic
- 1:02:58questions. And we'll be back in two weeks time
- 1:03:00with another stage talk. But for now, this recording
- 1:03:03will be up hopefully by tomorrow for you all
- 1:03:05to listen back to. And yeah, again, Claire, Elliot,
- 1:03:09thank you so much for coming in and chatting
- 1:03:10with us. Thank you so much. Thank you. I loved
- 1:03:13it. Thank you for listening to the stage talk.
- 1:03:17If you'd like to catch a stage tour live where
- 1:03:20you can ask the guest questions, join the Bellingcat
- 1:03:23Discord server by visiting www .discord .gg slash
- 1:03:30Bellingcat. The music you've heard is titled
- 1:03:33Dawn by Newer Self and is courtesy of Artlist.