Hacking Fake News: How Hackers Can Help Fact Checkers
12:55AM Jul 29, 2020
Welcome back to the hope 2020 live stream and I'm hoping you're all caffeinated This next one we have Christopher guests with us, who is from the Duke University reporters lab. For those of you not familiar with that it is a research organization at Duke University that is focused on journalism and public policy. And he is here to talk to us today about how we, and how hackers generally can help out the fact checkers. Over to you, Christopher.
Good morning, afternoon, evening, whatever it is, wherever you are, and thanks for tuning into this talk that I'm about to give hacking fake news, how hackers can help the fact checkers. So, first question is probably Who am I? Well, my name is Christopher guest's on the lead technologist of the Duke reporters lab at Duke University. We're part of the School of Public Policy. Most of my team is down in Durham, North Carolina. However, I'm lucky enough to be based here in Brooklyn, New York. I'm really disappointed Actually, this. This conference when it was in person was actually close enough for me to bike over to, which is a real shame. But I'm sure we'll all get through it. And I'm really excited to be just talking to you all this evening. So a bit of the background on me. I've been in the world of journalism for on and off for about 10 years now. I've been with Duke plus a bunch of other organizations working solely on journalism, and computer science for the last five. Over the last 15 years, I've sort of also been a programmer. I've worked in Eastern and Western Europe. I've worked in East Africa working across the United States. And what do we do at Duke? So what is my job? I we partner with fact checkers, we partner with media companies around the world, and we look to improve factchecking we facilitate policy conversation. We propose solutions, we research new techniques, we do all the background work that a lot of organizations simply don't have the time, money or expertise to really work into. And so we help. Now to get to the start of the actual content of this conversation, I'm going to first begin with the idea of trust. And to define this word, it's a willingness for an audience to believe to vote for and to act on what a public figure and institution or really any other person says. So what whether it's somebody is telling you to do something to believe something that is all around trust, right? And so trust is not a bug in the system. forcing people being able to trust things is really, really important. It's what makes us human. It's what society is. It's a social contract between multiple parties. Not to individuals only, but across everybody that they come in contact with and everything that affects. Now, it may seem like if if all trust is going away, but this is this isn't a new phenomenon. This has been forever, right? Fake News is not new. I could definitely give some examples about Thomas Jefferson and john adams in the early 1800s that are honestly a little too body even for this community in this conversation right now. They get they get pretty nasty, even, you know, 200 years ago. And, you know, trust is still here. It might not be at the same level as it was when we were younger, or when our parents were coming up, but it's definitely still here. It's just that it's over the last maybe 10 or 15 years, gotten much more of a base level. There's less trust in experts. There's less Trust in the government's people tend to instead trust people that they already agree with. And that can lead to a lot of different phenomenons including the ones that we've all heard of cuts such as filter bubbles, but there is still trust, we just need to figure out where it's coming from, and what the reason behind it is. Um, but to continue with the idea of sort of, you know, setting the stage, I'm going to go over what the basics of fake news are. So, first, I want to give a caveat, which is that us in the factchecking world, we tend not to use the term fake news. It's a little too broad. It's a sort of been co opted by certain political groups and politicians in the US and in the UK, and basically anywhere it's been translated as well. Somebody is using it to further that Their own particular political agenda. And so we we tend to tend to use a couple different other phrases.
The two big ones are being misinformation and disinformation. And it sounds sort of similar at the beginning. But let me just go through the quick differences. The main one being that misinformation is not necessarily malicious, right? It's wrong info that's spreader believes it to be true. So the person posting and the person talking about, they believe that what they're saying is actually accurate. That is not the case. For this information, though. This information is specifically to afford usually a political agenda, or some other social change that they wish but they do it through knowingly false information. Now, this information is often spread with the goal of it becoming minute misinformation that if people repeat the claim it'll be assumed to be true. And this is, this is dangerous. But we do have to consider the differences when we're talking about who we're addressing. What are the remedies for what version of fake news that we're talking about?
so let's look at the types of fake news that is usually spread out there and the types of misinformation disinformation. So the one of the caveat, one of the things that people don't usually realize is that when it's well done, disinformation campaigns are not usually obvious. Sort of like the adage of you know, I can always spot a bad toupee. Well, you can spot the bad ones, but you don't know the ones that you don't see. And this is true with disinformation and misinformation as well including headlines, everybody thinks, oh, I can tell misinformation from my headline. And that's usually true, you know, but the good disinformation campaigns don't do headlines like this, you know, this is Taylor Swift saying we should remove the Statue of Liberty and how can we have a monument to freedom that was built by slaves? Um, well, I wasn't really built by slaves, as far as I'm personally aware, is built by France and put together a New York Harbor. You know, well after slavery had ended, so that's definitely not the case, right? This is something if you remember, at all high school history, you're probably going to be able to pick up on in a second. But this is not the norm. This is not the ones that actually spread and get affected. Instead, the ones that really get out there are conceivable, they can be real, it feels like they could be accurate, right? Things like this one. The Fast Food CEO donates $400,000 to Trump The idea being that the CEO of yum food company that owns Wendy's and they own Taco Bell and they own Pizza Hut, donated this huge amount of money to the to the Trump campaign feels like it could be accurate big business guy gives money to hardcore conservative, you know, hopefully taxes go down. That that's possible. It turns out it's sort of true. Instead of it being the CEO of the entire brand, it's actually a fairly moderately sized franchisor, mostly in the American Southwest Arizona, West Texas, those areas who did give this money. So if you're in Arizona, and that goes against your politics, maybe you should consider a boycott. But if you're sitting in Montana, South Dakota or Portland, Oregon, it really has nothing to do with it. It was a totally different organization. And so, like I said, these are conceivable and going off that topic. You can go even deeper. So, for subjects that you're not experienced in, you trust other people not to lie to you, you trust that when somebody says they're an expert, they're actually qualified. So it's easy to assume that when, you know, a statement or a claim is made by somebody who purports to be an expert, that it's true. But that's not the case, necessarily, right? People can lie, people do lie. But unless you do the research, unless you look up the biography of that person, at least on Wikipedia, at a minimum, it's going to seem perhaps, that they could be telling the truth, and it's really easy to share that information I've personally found for this one before. Here's a great example. So Richard, oh, Jetta, or hetta. I'm not exactly sure how to pronounce his last name, but he's a democratic politician. And he wrote this the statement I'm loving watching these videos. Senators and members of Congress saying the same old stuff concerning gun violence and listening to them get their asses handed to them by citizens that are tired of seeing 75 to 100 folks killed a month and over 200 wounded by senseless gun violence. Okay, well, the fact that that is a massive run on sentence aside, this is a this seems like something that could be true. This you know, if you're against gun violence, this feels like it could be accurate, this feels like something you should care about.
fact, it's not accurate. PolitiFact rated this mostly false. And the reason is because it's actually too small. Turns out these numbers are more accurate to the daily attacks of gun violence than it is on a monthly basis. And you know, if you're conservative, you might take this and run with it. If you're a Democrat. You might take this and run with it, but it's still miss accurate, despite him purporting to be and to be an expert on this. And the fourth way that this can often happen, and this is this is particularly malicious is context, it can often be the wrong context. So the photo, the image, the quote, could be entirely accurate. It's not photoshopped, it's not copied, but it's just just represented wrong. And here is a great example of this. So, if you read this, this woman Allison, by the way, please if you have Facebook, make your account private because this is very public. So that's why I'm showing it. This is published on PolitiFact website, anybody can get this so I'm not particularly worried, but please make it public. Rather, please make your Facebook accounts private, please. So here Allison is reporting that these crates of rocks are dropped off by George Soros and LA is applies for looters and writers right. You know, back in May when all the rioting and looting was coming out. So this was actually posted on June I'm assuming that the photos were probably taken a day or two earlier. Putting aside the mildly implied anti semitism around George Soros being the one that's causing this, and the fact that staging Riot supplies and broad daylight is pretty categorically insane. Um, these photos are real. This is an actual photo. Nothing is Doctor nothing's cropped. However, the rest of everything that's been written is completely wrong. Instead of these being crates of rocks being dropped off for writers in LA, it's actually a photo of barriers protecting a synagogue in Oakland, California, you know, four or 500 miles north of there, and really the exact opposite of the statement that she was making. Um, and, you know, this isn't the only thing right? So I'm sure you See the, you know, about 10 years ago or so, you saw these people claiming that on September 11, as World Trade Center was being attacked that Muslims were celebrated in New Jersey and there's videos that it actually turned out, yes, they were Muslims, but they're actually celebrating a soccer game. And, uh, when in Germany about 10 years even earlier, and that's spread like wildfire because you, it's hard to find the original source these things are challenging, right. So that is it, you know, context is king in a lot of these situations. And then the fifth one is one that we all sort of assume happens all the time anyways. And that's just that just just played made up photos. Um, this is a great one. This is one of my favorites. It's been making the rounds for at least a decade. It has at various times been New York during Hurricane Sandy Houston during Hurricane Harvey in Puerto Rico during Hurricane Irene and New York during Hurricane Irene. I've definitely seen it and purported to be various cities in Southeast Asia during typhoon season so Manila during hi i and i think Myanmar, but don't quote me on that part. Anyways, uh, the thing is that it's not even bad context. This is the actual source of part of the image. And the rest of it is just photoshopped in. Whoops. By the way, thank you for Snopes, Snopes figured this out in 2011. These are the same exact quote is still going around everywhere. I see it pretty much every hurricane season lookout for it. You'll see it this year. I guarantee you come You know, late August early September. Okay, so given the subtleties How does factchecking actually work? Right? What is the process to discover this to discover this system? Okay, well, let's go next. So what I'm going to do is I'm going to take you step by step through this process as the journalists do it. And hopefully, you'll be able to get some insight as to how the factchecking side of this works, and perhaps how this could be hacked, how this can be modified, how this can be improved, how this can be automated, and feel free to, you know, come up with some ideas and bring it up here in the q&a.
So what is factchecking? Well, it's actually two types of backtracking. So just to be clear, one is there's internal fact checking this type has been going around since the beginning of publishing itself. Basically, it means going through an article or a book or something before publication and verifying the accuracy. So calling sources to verify the quotes are accurate. Looking up statistics, making sure people We're where they said they were when they said they were that the there's not plagiarism going on on the side of the author, etc. It's essentially that type of fact checking that we wish politicians would do themselves or your, you know, racist uncle would do before he starts posting stuff all over Twitter. But the second version of this is what I call external factchecking, which is holding politicians and public actors accountable to what they publish or speak themselves. It can also nowadays include verify and publicly spread media so means posts on Facebook and everything, but the original version of this was purely based on politics, and this has been around since about 2007. When my boss at Duke bill Adair created PolitiFact while he was a reporter at the Tampa Bay Times. And in response to the 2008 presidential campaign, they originally thought that they were only going to do it for the money. campaign
So what is this
process? Well, first fact checkers have to find the claims. This is actually sort of the easy part. They do it by watching TV by reading transcript scraping Twitter. There's a program wherein Facebook will actually help surface claims that have been reported and literally paid the fact checkers to look into that side don't Facebook does not pay them based on what they determined. There's, there's no none of that, at least as far as I'm aware. And I am not part of that program. So I've taken no money from Facebook for that part of the program. They have donated a little bit of money to the Duke program, but they hold no influence over what we research or what we work on or anything like that. Just full disclosure there. So second, once they have all these statements, they have to actually find what are called fact checkable statements right. So So this can be challenging unfortunately. For instance, someone's opinion say the EU is responsible for the immigration issues in Britain
is not really fair trackable, that claim is vague. And it's far too open for interpretation. Nobody really knows what the politician is exactly saying in that situation. And so trying to go down that rabbit hole will just eat up your entire week. Instead, facts, figures, historical claims, these are the sorts of things that that chick look for, as they're
trying to find, you know, statements that can actually be reported on.
So one of the funny things is When I do this sort of presentation as a workshop, I actually hand out transcripts of depending on what continent I'm on, like Prime Minister's question time in the UK or the State of the Union in the US. And I asked participants to actually go through and highlight fact checkable claims and then afterwards, I asked them, How would you actually research that claim? And it's pretty much a given that every person underestimates how hard this is and comes back with mostly opinion articles. Okay, so next, the fact checker needs to figure out where the speaker is getting at. statements can mean a lot of things to a lot of people. And so the fact checker has to figure out which angle they want to sort of attack this from is a politician talking about, you know, gun reform, or healthcare reform, if they're talking about suicide, or about suicide deaths, right? Is that mental health angle or is that a gun angle? It can really vary and those are different Toriel decisions that have to be made by the factor themselves. And then once we have these statements, we know what we're going to be looking at. This is where we get into just plain journalism. This is where reporting starts to really happen. Usually, it starts with the reporter just calling the Speaker of the claim and asked what their sources for that claim were. So you know, if somebody stands up in Congress or parliament and makes a statement about the, you know, percentage of solar production in the next year, they'll usually just call the politicians office, talk to a staffer and be like, Hi, do you know where your boss got that information from? And you'd be surprised how often that actually works? Like, like, a lot of the time that's usually like, where it can stop and you know, about an hour you can pound out a factor um, but You know, some people like to just make stuff up in the case of the previous claims they were looking at, you know, if something was fabricated or something that was just taken completely out of context, they might not be so willing to really answer your questions. In which case, it's time to go deeper. That can include, you know, interviewing subject matter experts, they you can start getting even deeper in finding the stet the statistics yourself scouring court records, if it's about, you know, so and so got arrested, public statements that people have made before, you know, is this something that has been claimed by the same person in different forms? Is it about something somebody else claimed? Well, then you got to go and you got to look at what you know, what area were they talking about? Whatever is subject, when was this made? Was this on the record? Was this off the record? You just have to sort of churn through all of your sources. Okay, and You know, once that's all done, once you've got through the whole process, you have to come up with the conclusion. Now, these can be positive or negative. Some organizations like to use ratings. So if you're, for instance PolitiFact, or the Washington Post, you'll use the, you know, Pants on Fire rating or the truth of meters, as Bill pointed, or you'll, if you're the washington post will use Pinocchio's or pedos. checkmark. If your organization such as again to go back to the UK, full fact, in the UK, or factcheck.org. In the US, instead of writing something directly on a scale, it'll often be more of a analysis Did you know going through step by step and point by point refuting or accepting what was being said or discussing those sort of vagaries around it, and both of them are acceptable. Each organization has their own sort of Their own sort of editorial policy on this. But it is a long process. And it's usually a long debate among the different editors to decide exactly where does it fall on the line. And then finally, you publish your report your findings, you publish on the web, you submit to, you know, Facebook, if you're in their programs so that they can automatically tag the posts that you were looking at, or Google for making sure that your factcheck gets included in various search results, which most of you have probably seen that that's a huge that was one of the big one of the big gets for the whole program was getting Google to actually put that information out in front of people.
But once that's all done, seems easy. It's not always that So what are some of the gotchas? What are some of the issues? Um, well, as I've shown fact, checks are often more than figures right? their opinions, their historical situations, their mislabeling, they're made up. And so as I mentioned in the previous slide, you have to dig into this, this takes time, this takes effort, this is not very automatable. Because just figures, you can sort of go through known information and just look it up, you know, based on various dead statistics. But that's not really possible if somebody is, you know, making more of a judgment call. And that's, in addition to the fact that politicians are professional obfuscators, as I like to say, they, their job is to make sure that everybody in the crowd listening or everybody you know, at home listening, or everybody on the internet reading, gets what they want from the speech doesn't have to necessarily mean what the politician meant specifically, but everybody sort of has to come up with their own ideal version of what was being said. We see this all the time. Right. And that particularly gets into a complicated aspect because of, unfortunately, the English language in that there's double entendres. There's word order vagaries, there's multiple definitions. It makes it really easy to say 12 things while saying the 13th thing. And that is a huge issue with, especially English. It's my main language that I focus on. But other languages definitely have this from English is particularly complex, however.
And then, you know,
context again, that's gonna come up a lot here. But context is so important when it comes to trying to fact check. So, say a politician says the tragedy this morning. Well,
you have to
know where you are. Right? Because it could be that could be the tragedy of you. Don't something that happened the next coming over something that happened across the country, something that happened in a different country that morning. You have to know what day it is because what is this morning? Is it you know, was it 11am? Or was it 8am? Did the event, say an event happens at ADM event happened at 11am? The speech happened at 10am. But we're checking at 3pm. You have to understand what what context that the speech was given? And who is the audience, right. If the tragedy this morning was that the Supreme Court ruled to block abortion rights, then it's only a tragedy to a specific subset of people that you're talking to. While it's actually a huge win for another subset, not passing judgment on that at the moment, but you have to consider that. So what is the frame of reference that the speaker is speaking from is really really important and very, very difficult also, at the same time, especially if you're trying to do computer computing or natural language processing or anything like that. Very, very complicated. So, now that we've got through that whole process, where do you come in? Well, first off, I want to say thank you for letting me speak today and that we welcome the hackers, please come to our side. A bit about the community that I hope you all join. There are 78 International fact checking network certified fact checking groups around the world. The ifcn is a group that creates a statement of purpose and a set of ethics and a set of guidances and then certifies organizations. And there's,
We have a conference every year. It's very fun. But there are spots 60,000 facts that have been done by fact checkers around the world. That's quite a lot in my personal opinion and the data Base grows literally every day, just this year in the database, I personally run for our partners at the Washington Post PolitiFact and factcheck.org. There's been, you know, 2749 as of the time of recording this conversation, and so by the time that you all have are watching, it will definitely probably be over 2800. And that's again, just this year, that's not even talking about the years past, that's just in the last six months. So why we need you is that most and really pretty much almost all of the fact checking organizations out there in the world, they don't have tech teams, they don't have coders or programmers at most, they have you know, maybe someone that helps the website and runs a WordPress, um, but in general, they have no technical support. This is all done manually, all done by hand and it's It's a lot of work. So you know, I mean, just the percentage of social media reports that are checked are vanishingly small that are reported. I can't give exact numbers from Facebook, but just know that it's closer to single digits than it is to double digits on the number of reports versus the number of things that are can actually be checked by our fact checkers.
misinformation is malware. This is my theory.
In that it is something that the system was not designed to prevent. And like the internet, much of society is dependent on the idea of trust that I was describing earlier. And misinformation, exploits that trust, it takes over, it seeps in and then it spreads. You know, I've said it before that there are more similarities with how a Windows exploit travels across the internet. In how it digs its way into a system and how it recreates itself, then there's more similarities with that and misinformation, then there really are differences when you look at from more of a conceptual level. And so from there, given that, let's assume that malware misinformation is a form of malware, how do we harden that system? Where do we look for the exploits? What are the topics and concurrent events that misinformation are taking advantage of, to seep their way into our society, into our culture, into our politics, into our everyday lives into our relationships with each other? And with our families? We've all had these things, you know, I'm sure we've all had struggles with this over the last couple years, and especially over the last four or five months. So what can we do as people who think logically who think systemically about this are their patterns of who starts in misinformation who shares misery Information what causes people to share information? Is there some sort of fingerprint we can develop to ID this misinformation malware as it spreads around the internet and through society because it's pretty much all spread digitally these days, especially during lockdown system during lockdown times when seeing another person in person is not really an option. So what can we do to try to identify this information? At that point, it's what I call system social engineering at scale. We all know about social engineering we all grew up with stories of Kevin Mitnick but what can we do to take this and operate it at scale? So, what to citizens? What do social media organizations What do you know reporters? What do news organizations What do we do to convince people not to spread this information in the first place? It's sexy to spread misinformation it gives you power it makes it seem like you know something
which you know,
is very important for a lot of people, especially days like today, where there's not a lot of self governance over what you can do on a day to day basis, we're all stuck inside, we can't really move. So spreading this information, you know, acting like you know, something more than others this this is, you know, this is psychologically very important. So, what can we, as people who believe in truth believe in validity? What can we do to help this you know, once it and you know, be if we can't stop it, what can we do to maybe somehow put the genie back in the bottle just a little bit, right? can we can we tamper it down? Can we stuff it away? Can we, at least, you know, know where it's going to go next? And can we teach people to look into this more? Can we teach them to be better consumers? Can we do this at scale? Can we do this at a scale that is not a generational scale? You We need to increase education in the schools. That's awesome. Yes, we do. But that's going to take 20 years. So what can we do in the next two years? What can we do in the next four years? What can we do in the next 10 years instead, to help the to help prevent the spread of misinformation, and this infection from going further? And at that point, what can we as hackers, uh, you know, do to address the ethical issues of this, this is something I'm sure all of you have already thought in your head like this, you know, how do we balance this question of fake news of freedom of speech, right is not saying you know, that that bag of rocks was put there is that not technically freedom of speech? If we stop that, is it technically censorship? You know, at what cost? Should this information be free is a question I would pose to that. You know, do we apply Western Silicon Valley Valley more or less rules to the world of this misinformation? And even then if we're going to talk about that, do we apply German ideas of free speech, or us versions of ideas of free speech? Because if you take a photo of a Nazi rally, and you spread it around saying that this is, you know, these are the, these are good people, or these are bad people doesn't matter. In Germany, that's often illegal. In the US, of course, it's it's very much protected speech. So, which ones do we implement? Do we remove that completely? Do mark it as false? Do we, you know, just put an overlay like Facebook tends to do or do we delete it from the internet is German law sometimes requires you to do for that matter, you know, who are you, me or any of us to actually determine this validity who deputized us who watches the watchmen I guess probably Probably the cliche way of putting that, but it's true. Uh, you know, the way that most factory workers and most researchers like myself tend to mitigate that is by making our process very open, you know, we start taste that I the many of the ideas of the open source community and make it available and open and all of our donors are listed, etc. But do we have to? Is there a better way to do it? Is there more that we could do? You know, in some place that can often be very, very dangerous if you're in countries with a lot of political suppression, saying exactly where you get your funding from? can lead can leave you in jail. You know, taking money from George Soros in parts of eastern Europe or Russia can literally get you killed. But we still have to trust them. So how much do we force think these are very quick, very, very important questions. And when something does go wrong, when something is missed, who gets the blame? Is it you know, the fact checker for Not getting their facts right? Or for missing a completely. What? Where do we go with this? These are questions that I think hackers have been thinking about for most, you know, ever since sort of our field of interest was created back in the 40s and 50s. And so having that consciousness and having those discussions and having the viewpoints of this community in that is going to be really, really important, especially as we go towards more technical solutions.
And that being said, What can you do else? Okay.
So what can you
do? Well, this is sort of a laundry list for me, but it's a few places to start. And there's way more than this slide gives us but number one, you know, design algorithms to track claims from origin to mass acceptance. This can be used to figure out who's benefiting from the spread of a specific falsehood. Who has what to gain And how can we hedge it all off at the pass? Right? Next is claims are repeated, they're repeated often. Is it possible to compare them to something that's already been checked? This is my specific field of research at the moment is trying to take place somebody is saying off the TV, or in a speech or in a newspaper or on a website, and look through our own database and see if somebody else has made the same claim. The hard part is it's usually different wording, it's different phraseology. This is where the English language becomes way more complex than say, doing the same thing, as my colleagues in Argentina are doing with Spanish not saying their job is easy. But you know, it's it's, it's just a little bit easier. Um, so, perhaps we can automate the tagging of claims as misinformed. Maybe we can even stop it and track that way. You know, if something comes up that's already been stated in a different phrase. We can put that tag on it right away. Um, can we hold tech companies to standards. This is a hacker conference, right? Large companies being bad is usually a fairly acceptable stance to take here. But aside from advocating for the fediverse, what can we do to keep it honest? Um, you know, if we work at these companies and if we're in meetings about misinformation, can we make sure that the threat of this information or how it spread is addressed when discussing product from discussing policy? If you work at one of the large, large companies, right? Are you complicit in your, in the spread of some of this? Do you, you know, sort of ignore it? Do you sort of brush it off? Because it's not a priority for your manager? Can you insist on making it more important? Can you insist on making this more of a priority than just making sure that you hit the time deadlines for you know the next q2 release? And can we come up with novel ways to examine these social ideas? From a technical point of view, you know, can we bridge society, sociology and graph theory? Can we come up with models and algorithms to search these areas? I know that there's been work on this in various academic universities and in the fields, but that stuff is not penetrated within the factchecking community, we don't really know much about it. We can't really use it. So how can we bridge that gap? So if you're working on some of this, please, you know, please do let me know because I'd love to bring you on to some of our conversations. Um, on that note, I want to thank everybody for tuning in, especially those of you a little bit east of me here in New York. I know it's late and I hope that your sleep last was well spent, and that you all learned something. I want to also thank HOPE for transitioning so well, for this new format. I'm sure we'll be all sipping mai tais at 4am sooner rather than later. You know, the news is bleak everywhere, a little bit at this moment, but I think we'll all get through it. And lastly, please stick around for that. So hopefully
you enjoy Christopher's video. Christopher, welcome to hope. And we have a few questions lined up in the matrix chat for you. I think we got quite a few questions already. But if you have a last minute question, anyone want to stick it in there? please throw it in the q&a there. So first of all, if factcheck fake news can still get repeated for a decade.
cognitive biases can cause people to just double down on misinformation is fact checking itself a losing firefight? What can we do to take it to the next level?
Yeah, that's that's a great question. And it's one that keeps pretty much everybody in the factchecking. World up most nights. Yes, some people double down. Not most people as far as I've been able to see from To publish studies that have been done out there, so there is room for correcting the record correcting the record alone is just important, right? You know, people can say things, but it's important to have an actual history of what is the truth available to the public. And, you know, there's a lot of different methodologies for how you can sort of intervene where somebody is receiving the information, the sooner that you correct it in their mind, the less time it will have to sit in. And we do need to do more. We need to work on that intervention work on getting information corrected sooner quicker, more immediately in front of people on their TVs in such a way before. It does have a chance to sort of set it in those cognitive biases have a chance to grow.
Thank you. How can consumers train themselves to detect misinformation themselves?
Oh, great question. Great question. Um, the answer is and I think It's one that this crowd will appreciate question everything. Um, you know, basically, there's an old newspaper adage that if your mother tells you that she loves you check it out. And it's something that we sort of need to get around to, in the world to you know, don't believe that retweet just because it's somebody famous, right? Again, this crowd is probably not who we're speaking to, but perhaps you know, some of our more elderly family members, for instance, um, make sure that you don't personally spread information, check it out, and make sure that you have the facts yourself, before you've spread something. I've been guilty of spreading stuff that I, you know, got lazy and didn't check. And it was proven wrong later. Now does everybody but if you can stay a bit on top of it, you know, just do a quick little search and see if any other places are reporting on it, that sort of thing. You're going to, you're going to go a lot further. Thank you.
We had a couple of questions along the lines of Chris custodia tips. custodies and with the one that I think was more interesting, went on to ask, What if the fact checker groups conclude something different between themselves? How do you address that? Do you address that? And can that be used as false confirmation by certain interest groups? Oh, yeah.
Another great question, one that I actually don't think I've gotten before so but something that we've thought about. So, you know, the facts records themselves are independent entities, right? So they can come to different conclusions, and they can write their stories. And if you read them, they'll step you through point by point how they got to those. And so you can, you know, you can go in and look for yourself and see which methodology you prefer, you know, maybe it's something in the middle that actually works better. The fact checkers, in my experience tend to come to pretty similar conclusions, like maybe they'll say like false or mostly false but in general tend to be pretty even keel on that. But when it happens, you know, we don't you know, it's not like the Washington Post calls a PolitiFact. It's like, why did you do that? Nothing like that happens, right? We just sort of move on with it. And I'm sorry, I forgot the last part of the question real quickly. Yeah.
I think they were saying that maybe if that occurs, it can also end up being used as false confirmation. Ah,
Sure. Sure. Sure. Sure. Sure. So, uh, yeah, that's true. Um, unfortunately, you know, with a free press and free speech in the US and most other countries that I'm sure everyone on here listening to from hails from, you know, that's, that's just the thing. If you publish something, people can twist your words you can potentially sue for libel. But there's, you know, there's really nothing you could do, what
effective versus implications for you and how it is used and how you use it.
Um, so, the fediverse um, I am pretty sure that I am one of maybe three people in the entire community in the world that's even heard the word fediverse I'm on mastodon. I'm at sea guess I'm on the, but, um, in general, we haven't touched it. Um, I actually have a ticket in GitHub right now on one of our products to actually be posting or fact checks to mastodon, but it's just not a priority. Um, you know, it. I'm not 100% sure on this, and this is just my opinion, but I imagine it's going to be a little bit like preaching to the choir, like the people that are on fed verses are probably more likely to be the people that Google when they hear, you know, that do their own research and are less likely to fall for it. But that's just that's just my guess. And you know, maybe I think we're all better than everybody else or something.
Fair enough. So I'm going to ask you You this one, what if the other side is legitimately different opinion and not facts are not fake news?
Yeah, a good question. In those cases, fact checkers tend to not fact check that. And one of the initial sort of selection criteria that fact checkers go through, as I mentioned in the talk is sort of weeding out things that are fact checkable items that are that deal with, with truth and numbers or events or what have you. And not opinion based, if it's opinion based, we tend to sort of we tend to just let that go and not not address it.
So is hydroxychloroquine working fake news.
Oh, I think there's probably a comma that needs to be in there somewhere because this could empty different definitions. Another problem with backtracking By the way, um,
Don't know. Um, according to the
FDA, yes, I believe because they pulled their emergency at least last I saw they pulled their emergency authorization to use it on COVID. But I, that's just me reading the New York Times. Um, I'm sure one of our factors has worked on it. You know,
and I think that that kind of plays back the last question. That was a question that was posted to the matrix chat, but but I think it plays back to the last question. And as much as when we're looking at the news, how do we distinguish between what is fact checking and what is opinion? I was tempted to say what is fact checking and what is journalism but I suspect you would consider them you know, what one to be a subset of the other correct?
Yeah, yeah, I would. Um, so how do you how do you do that? Well,
you know, they can look somewhat similar. If you're in a newspaper. Hopefully the opinions are in the opinion section a signed by the editorial board or a member of it in you know, legitimate news organized
But other than that, on other places,
look for sources look for the reasoning on how they came to that fact checkers, you know, especially the official, you know, the ones officialized by the International fact checking network and audited by them will, will step by step you step by step through the process, cite their sources directly and lead you to the end. So you can literally just tag along on that process opinions are going to be much more vague, they're going to you know, they're gonna use more outrageous or, you know, hyperbolic language and hopefully we try to stay away from that. Okay,
penultimate question for you now, huh? Are you familiar with one of the early designs of copy paste where in metadata was retained? Is this the type of idea you're suggesting when you request tracking claims from origin to access Since
a one not familiar with that, um, seems like something that would that would totally happen though. Yeah, that would that that could be an idea, right? This is the reason that I wanted to talk to everybody at this conferences, because there's a lot of great ideas like that, that I have not come up with that we have not come up with but that you guys might be able to.
Okay. And we do still have lots more questions pouring into the q&a channel. Thank you for that everyone. I'm not going to give Christopher the the question of what is truth or how does he define real, but I will throw you a lost bone and just ask this one real short mass acceptance or mass distribution,
fast acceptance or mass distribution. Um
that is it. Good one, and what I'm going to bring up at the bar next conference because this is going to be tricky to everyone, um, I would say mass. You know, for the hacker community, I want to say mass distribution, right? More information is better than less information. Um, I wish there was both I wish people would, you know, I wish the mass information was something people could mass accept. But, you know, I think more is better than less information in the world, personally.
And with that, thank you very much, Christopher. And that was a great session. And we hope everyone else will hang around through the break for our next talk. Thank you very much.