← Truth in the Barrel
Truth in the Barrel
Devil's Cut | Adventures In Misinformation w Renee DiResta
00:41:39
How much misinformation is being dumped on the Internet? Who is responsible and how can we better protect ourselves? Amy is joined by Professor Renee DiResta, author of the book, Invisible Rulers to answer all of these questions and more! ...
Loading summary
Transcript49 lines
- [00:01]
A
How did misinformation and propaganda get so bad? How do foreign governments manipulate public opinion? Is that happening here in the United States? And if so, what can be done to stop it? Hey, everyone. Welcome to the Devil's Cut edition of Truth in the barrel. I'm Amy McGrath, and today we're talking with Renee Diresta, who is one of the world's leading experts on misinformation, how it spread, how AI is affecting the spread. Renee is an associate research professor at Georgetown University, where she studies propaganda, influence and abuse in network systems. Before that, she led research at Stanford, studying everything from foreign interference in the United States elections to viral rumors about public health. She is also the author of a new book, Invisible the People who Turn Lies Into Reality, which looks at algorithms, influencers, and online communities and how all of those things combine to make certain ideas go viral, even if those ideas are 100% false. She's written for the Atlantic, New York Times, Wired, and has advised leaders in medicine, government, and tech. And when she's not talking about the future of our media ecosystem, she's also a mom of three, like me. So welcome Renee Diresta.
- [01:24]
B
Thank you for having me.
- [01:25]
A
I wanna jump right in. For a long time, Renee, newspapers, TV and radio were sort of the way that people followed politics. You know, you think of the Walter Cronkite. When I was young, it was Tom Brokaw and Peter Jennings. They set the stories, they shaped the conversation, they. They influenced what everyone was talking about. And people sort of made decisions and had debates based on the same sort of information, the same sort of facts. And then the Internet came right, came along and changed all the rules. And now you have these influencers, people call them smaller creators, algorithms behind the feeds that shape the way we see the world as much as traditional media ever did. And so I want to ask you about your new book, Invisible Rulers. You describe our information system today in our environment as a carnival of mirrors. Can you talk about that metaphor?
- [02:33]
B
Yeah. So I think that there's always a little bit of nostalgia for the old ecosystem where we had this sense that everybody was following the same person. There were a handful of channels. People were, roughly speaking, seeing the same information being presented. There have always been, though, these alternate media ecosystems, the term I write a little bit about the origin of zines back in the day, and counterculture publications and places where people went to try to find alternate information, particularly when they had the sense that the media might not be telling them the truth or it might not be representing a perspective that they held. And One of the reasons why this happened was, of course, because media had to appeal to what you might call more institutional advertisers. So this is the. You know, Noam Chomsky writes this book called Manufacturing Consent, and he describes what he calls the five filters by which media sometimes shapes its coverage to be palatable to these audiences, that it's trying to reach these mass audiences, and to the advertisers who want to reach these mass audiences. And he writes a little bit about. He's approaching this from a very sort of political leftist perspective. He writes about this from the standpoint about whose voices then are left out or whose stories don't get told, or perhaps more importantly, how are stories that do get told inflected in certain ways. And so this is a way to think about why, even when we were all seeing the same content, we weren't necessarily seeing 100% of the facts. So when you move into the media ecosystem that we're in now, you can no longer really appeal to the mass. To the mass public. The mass public has really fractured. You have. Niche publics is the term that I chose to go with, where you have different groups of people on different platforms paying attention to different creators. Part of this is because it is algorithms that are directing people to creators that they might like, to pieces of content that they might like. And the effect of this is that they're seeing content that is really produced by people who are representative of a political or identity based point of view, and they're speaking to audiences who they see as being just like them. Right. So that idea of this is a creator who is just like me. Instead of seeing somebody who is necessarily an authority figure or somebody with expertise or gravitas that appeals to the mass public, we've moved instead into this media ecosystem of niches with a very different set of incentives. They're not trying to appeal necessarily to mass advertisers, but they're trying to appeal to the specific niche that they are looking to reach.
- [05:22]
A
Yes. And that's the mirrors. You're sort of seeing something that looks like you reflected back.
- [05:28]
B
Exactly. So looking at something that is reflected back at you. And of course, there is also this element, you know, when we think of hall of Mirrors, it's a metaphor that comes from, you know, walking around, perhaps the carnival. And the mirrors aren't all showing you the accurate representation of yourself. Some of them are stretched out, some of them make you look wide. Right. Some of them are wavy. And so sometimes you are getting information that is really put through the lens of what the creator thinks you are going to want to hear, or the creator wants to hear for that matter. And so that component of it is that the information again is not necessarily moving in a way that is had the sort of verification and vetting that you used to see in some of the larger media outlets. It's instead moved much more into this opinion based, belief based content ecosystem where facts are maybe secondary to what is entertaining or what is appealing or what is likely to be receptive to that niche.
- [06:31]
A
Yeah, and this isn't new for foreign countries too. I mean, we sort of see, are kind of experiencing it here in the United States increasingly. So I'd say. But you've tracked disinformation from state actors like Russia and China. And are those two state actors, are they similar? Are they mostly doing this on Facebook, on TikTok, how does that work?
- [06:59]
B
So they go to whatever platform they think they need to be on to reach the niche audiences that they want. They also recognize that they're reaching a niche, which is very interesting. Russia really acutely realized this very, very early on, the very earliest stuff that we saw from Russia. So the data set that I analyzed for the Senate Intelligence Committee, which the platforms attributed, just to be clear, I didn't make the attribution. The platform said, these are the accounts that we found on Twitter, on Facebook and across Google's properties. And they went, and they said, these are the accounts that we've attributed to Russia. And they were all appealing to different niches. Right. So they didn't bother trying to create content that reached mass audiences. They recognized that they wanted to get into the weeds with different individual groups of people. That's partly because in Russia they wanted, the Russian trolls wanted to run a strategy of division. So the way that they did that was they pretended to be members of these disparate groups. So that might be a Texas secessionist page. There were about 30 different variants on the black experience in America. So there was a page dedicated to black liberation theology, a page dedicated to being black Baptist. There were pages related to hair and culture and music. And they would crib, they would actually steal from the Tumblrs in particular, but the content produced by real black women, and they would sort of twist it and put it out there again, hall of mirrors in their own way as these fake, these sort of fake black Persona accounts. But there were about, again 30 different niche pages and accounts that they had kind of carved out as representative of the black experience in America as they saw it. Right. They did this Listening tour kind of thing where they went all across America, they traveled, and they tried to get ideas for what they were going to do. And then this was what they came up with. And so they had maybe a dozen or so conservative pages. The older Reaganite conservatives got pictures of, you know, American flags and patriotism and amber waves of grain and stuff. And then the younger pages, the pages targeting younger conservative meme accounts, they were actually just grabbing Turning Point USA content and slapping their own logos on it. Just plagiarizing. Right.
- [09:36]
A
Who is the they, Renee?
- [09:38]
B
The Russian trolls. Yeah, yeah, yeah, I'm still talking about the Russian trolls.
- [09:42]
A
That's amazing.
- [09:43]
B
And the reason. The reason that they did this with the niches was recognizing that they could then pull people in based on their identity and then from there pit these identities against each other. So you would have a page dedicated to being for veterans. And again, they always spoke in the first person. As veterans, we. Or as veterans, I, as a veteran, I was the way that they would write these posts, which would appeal to the unity of the people. They were writing as if they were members of the community. So as veterans, we need our benefits. We are not getting our benefits. Why are our benefits going to refugees? Meanwhile, then they had a page called United Muslims where they were speaking as Muslims saying there wouldn't be so many refugees from Syria if the United States hadn't started a war. Now, keep in mind, Russia is fighting on the other side of the Syrian war. Right? And so you have the. So this is the incentive for them over there is also to shape perception about that conflict, which the US Is not paying attention to at all. The American public is not paying attention to it, but they are nonetheless doing this propaganda about the Syrian war. And the pages there are again saying why the United States is making these refugees and that it doesn't want them. Right. And so why does it hate Muslim refugees? Why does it that hate our people? And so you see these. These ways in which these identities are being pitted against each other, and that is the way in which they do it.
- [11:10]
A
And this still. This is happening every day. This is happening right now.
- [11:15]
B
Yes, absolutely. Yeah. And. And the, you know, the Russian trolls were interesting because back in. So from 2015 to 2017, late 2017, nobody was really looking for them. Right. So they grew these accounts, and some of them had several. Several hundred thousand followers. They were retwee people like Jack Dorsey. Right. People like. I believe, if I'm not mistaken, Sean Hannity, definitely. People like Charlie Kirk, you know, these people who were luminaries who were really influential themselves found this content compelling and would pick it up and would boost it. So I think that. And there were articles that CNN and others wrote about this model of who was boosting inadvertently, the Russian trolls. So they were definitely cribbing from Turning Point. And then you would see other right wing influencers going and boosting their content. I'm trying to remember if Charlie Kirk actually was one of the ones who retweeted them. This is not as a pejorative. It doesn't matter necessarily who specifically was the right wing influencers that were boosting. That was happening on the left too. It's that cnn, I think a couple times actually embedded their tweets in articles as like vox populi, like here's the opinion of the average guy on Twitter about this issue and it turned out to be a Russian troll. And you know, CNN owned up to that. Right? They published it. They too. So the point is, the point of this anecdote is really to relate that the content and material that they were putting out was compelling as being representative of that relationship. Like, this is somebody who is speaking as a fellow member of our community. Let's amplify them. And that was what began to happen after 2017, late 2017, when the Senate hearing happened and the platforms had taken all these accounts down, there was more of an effort to prevent them from growing again. And for a brief period of a couple years, that really became a priority of the tech platforms, particularly on Twitter, particularly on Facebook. You mentioned China. China also began to get in the game quite visibly, Funny enough, after the Russia stuff came down, you start to see China really moving in. In 2018 in particular is where the first sort of big Chinese networks begin to appear, targeting the U.S. china has run propaganda trolls targeting China for many years. Right. Targeting their own people. They begin to reach outward in 2018 or so. And this is when you start to see the difference in styles, the difference in tactics. They don't have these sophisticated Personas. They don't understand the niche dynamics of the American experience or identity experience. And so they, funny enough, are not really in the weeds in these different communities yet. They are still much more putting out a style of propaganda that is much more of a broadcast era style where they're actually doing quite a lot of talking about China and about Chinese issues. So they want to shift perception of the Uyghur, you know, the Uyghur problem. Right, right. The problem that America sees, you know, the. This sort of persecuted minority. The problem is from their perspective they write about it as the weaker problem. And so they, in order to address this problem as they see it, they begin to put out this content to really flood the zone with posts about how American political leaders are lying about the Uyghur problem. The Uyghurs are quite happy, actually. Look at these pictures of happy Uyghurs, you know, and so you start to see the. That is not being done granularly in the niches. It's maybe perhaps harder to relate the Uyghur struggle to different American political niches. But I mean, man, the Russians find a way. But in this particular case, you see a different approach to it, that they've gradually evolved over time as they've begun to realize that the tech platforms are not doing these kind of interventions to the same extent anymore. Academic institutions that used to do this work were forced to close. This was reframed as censorship as, you know, some vast cabal to find and pretend that like conservative perspectives need to be taken down. So as that, as that narrative took hold, ironically, domestically, we have made it much easier for these foreign actors to operate now.
- [15:35]
A
Yeah, it's like the propaganda of lies helped certain groups of people. And so, you know, they come in and they say that any type of movement to tamp that down, that propaganda is, you know, anti free speech or censorship and all that stuff. And I mean, do you think these Russian propagandists, maybe even Chinese are, are succeeding in dividing our country, making us less unified here?
- [16:04]
B
So they were always, you know, there were always fewer of them than of the actual domestic influencers that were doing this. And that's the thing that I try to tell the story in the book too, which is that they're real, right? They're unambiguously doing this. They're absolutely there for that period of time in 2015, 2018 or so, they are getting lift doing this. They are doing this effectively on that division front. But there is. Social media is noisy and it is very fragmented and it's hard to make a case that they become the thing that, for example, swings an election. Right, that, that becomes a big question. However, this question of are they there in the niches? Are they working to sow division? Are they there kind of, you know, I think about it a little bit as, you know, the chronic condition of making you, weakening your immune system, making you a little bit sicker, just having that sort of constant, that sort of constant poking. Yes, very much so. They're there. The question is, when I said Russia was pulling from Turning Point, content was plagiarizing turning point content. The reason they were doing that is because there was already a domestic voice in the conversation that was available to them that was producing the content that they wanted, that they saw as being useful from a politically identity based and division based perspective. So they had something there already that was useful to them. Which means that if you think about it, we are already doing this to ourselves. That identity based model of all Republicans are liars, all Democrats are liars, they want this, they want that. These sort of worst case caricatures that a lot of political influencers really thrive on. In the book I write about, Libs of TikTok and the phenomenon of nut picking, right, which is where you'll see somebody go and grab a random social media account, some random tiktoker barely has any followers, you know, maybe made a video that said something stupid and they'll go and they'll pick up that video and then they'll blast it out to all of their followers, like, look at what this, you know, lib or whatever thinks. And that becomes the avatar that their followers have for what a liberal or a Democrat or Republican or whatever it is believes. And so that model of sowing division, the influencer meanwhile profits, just to be clear, that's why they're doing it, right? They're doing it for profit, they're doing it for clout, and they're doing it for political ideology. And so the combination of those three things incentivizes the influencer to go and do this. But the net effect is that then other influencers see that this is potentially lucrative, that it's a way to get attention, that algorithms will reward it. And so they come in and they do it. And this is much more persuasive than any Russian troll, because again, the person who's doing this, the influencer who's doing this, is doing this with their face attached. They're forming that relationship, they're making themselves part of a community, they're speaking as part of a community. And the Russians really could never do that. They pretended to be part of a community. I think with AI, you're starting to see a real resurgence of this where they can put a much more convincing face on it. But they didn't have that last piece available to them. And that's why you saw them eventually just move into paying influencers like Tim Pool, like Benny Johnson, like Lauren Southern, rather than trying to kind of to do it themselves.
- [19:39]
A
It seems so overwhelming when you talk about it. I mean, it just seems like we're never going to be able to break through this. And a lot of people who are listening to the show are probably wondering, oh, my gosh, what can we do as individuals, as a country? Is there anything. Well, let's just start as individuals. Is there anything that the average social media user can do to sort of resist being pulled into one of these craziness? I mean, it's everywhere.
- [20:11]
B
So I think one thing is recognizing it when it's happening, right? That the people who make their money by pitting different groups of Americans against each other. I think there has to be a bit of a social shift against that, a norm shift against that, where people begin to think like, this is tiresome. This is not something that we want to see every day. It's just, you know, it's just not what I want in my feed. And then to be muting those accounts rather than engaging with them, to be, you know, you don't have to block them, obviously, but you can just hit the mute button and then you don't see them in your feed anymore. That takes away some of their attention. That takes away their ability to earn money. And then you're shifting the incentive, right? Or you're making it so that you're making it maybe a bit clearer that that is not going to be the most lucrative path going forward. But that has to happen at sufficient scale that they begin to realize that their behavior is being penalized by a potential audience rather than rewarded by it. So that's a bit of a norm shift and a social shift on social media platforms themselves. I think that's where you start to get at questions like, can we make them more transparent again? There too, there's the question of can you give users more control, which would let them maybe turn off some of this kind of content. Just say, I don't want to see certain types of content. I want to see. I want my feed to prioritize certain types of topics, not other types of topics. You see Instagram starting around the edges to look at this, right? Recognizing that it does shove people into rabbit holes and maybe they want to get back out of it. And so giving people an option to say, you know what? I'm really tired of getting this kind of content in. The example that came out with this announcement from Instagram is two days ago, three days ago. One of the things that was sort of funny was the example that the reporter used when he was writing about it showed radical politics as. And we were all kind of wondering, what radical politics. What is Instagra? Think that term means is that considering.
- [22:21]
A
Everybody labels everybody on the other side as radical.
- [22:23]
B
It was sort of a, sort of a funny example. But. But this is a, a thing that you're just starting to see platforms beginning.
- [22:32]
A
To ask, is that called middleware? Is that what that is?
- [22:36]
B
Middleware is something that I write a lot about. In this particular case, it's Instagram surfacing a piece of its, its own algorithm to its users. So there's no third party there. Middleware refers to third parties that can come in and produce code that acts as an intermediary between a user and a platform. Hence the middle part. So when we talk about middleware, we are talking about things more like there was this product called Block Party that was really great, and blockparty was a tool that you could use. You paid to subscribe to Blockparty and high volume accounts used Blockparty. I used Blockparty even though I only had about 80,000 followers on Twitter. What you would do with Block Party is you could use it to mass block if you wanted to. Where, if you dealt with, you know, if you got piled on or you said something and, you know, some members and faction of the Internet got mad at you that day, you could just hit a button and it would. And it would block people who had engaged with a particular hate tweet. So something that happens quite often. I'll use a specific example to be illustrative. Jack Posavec accused me of censoring Hunter Biden's laptop. Complete bs. Out of the, out of nowhere, just nothing. There was absolutely nothing at all that it was made up. Totally, 100% made up. He just said it because sometimes they just say things, right? And he was mad about Elon favorably tweeting something I had done. And he was like, oh, that's the woman who censored Hunter Biden's laptop. And I was like, I didn't work at Twitter. And I said, I thought it was a bad call. This is complete bullshit. But nonetheless, 5000 people were all of a sudden in my DMs screaming at me because of Jack Posavec doing this, right? And the sort of so Block Party was useful for was being able to hit the button and say, anybody who had engaged with that tweet, I could simply block or mute for that period of, you know, for in that moment, which is very useful. So a tool like that just allows the, the app to act on your behalf with the platform. There are other ways in which. So Middleware can be used for moderation Giving users more control of their moderation experience. This is not to say, look, I could, what am I going to do otherwise? Like, you go and you hit the report button on a tweet and like, there's nothing.
- [24:59]
A
Nothing happens.
- [25:00]
B
There's nothing. Well, because sometimes there's nothing that they do that is technically harassment. He said a thing, it was a lie, but it's not going to get moderated. So all of a sudden the onus is on me to deal with it. And so in this particular case, middleware can give users empowerment tools. It doesn't solve the problem, which is that this person was incentivized to lie and he made money off of that lie. That's the other problem that we need to talk about, which is what do you do to shift those incentives? But it did give me some capacity in that moment to change my experience and to deal with the immediate situation. Middleware. Another area where it's very useful is curation, is saying, you know what? I'm really tired of seeing AI slow slop. I just don't want to see AI slop. I'm tired of the proliferation of AI generated images all over my feed on Facebook. Wouldn't it be nice to have something, a third party tool that could help me control my feed better, give me some granularity in a platform that otherwise doesn't. Blue sky is fantastic. There's a lot of granularity there. It's. I mean, it's fantastic in the, in this sense. I know a lot of people feel like it also has problems with being very insular and very, you know, the tone can be negative for certain groups of people who are there. But also you have that power to use what it calls labelers and feed creators to really directly shape your experience and to, you know, to block and to hide and to subscribe to shared block lists and things like that. People who might be toxic in your definition of that term, or to create feeds that show you content that you want to see again with your personal preferences in mind. So that's what middleware is. It refers much more to having that third party thing that operates on your behalf to make a platform behave a little bit more in a way that you would like it to. And you pay the third party provider, in this case, where you subscribe to the third party provider, that does that for you.
- [26:59]
A
Oh my gosh. And it seems so overwhelming for somebody who's working every day, not, not they're just on social media because, you know, we're all on social media. But to know how to do those types of things that you're just talking about, it just seems a little overwhelming.
- [27:17]
B
That. And that is the. That is the question, right? How do you make it so simple? Because middleware would have to be simple in order for it to be widely adopted. And even those of us who study it are very well aware of that. What we see on bluesky, where it's the most readily available, is that a lot of the times the users don't really know what it's capable of or what it's there for. What this translates into is that they often then keep trying to appeal to centralized moderators to moderate in a way that they want, and then they get frustrated when the moderators don't do that. And so explaining a little bit more how a labeler can really very directly achieve the results that you want immediately in that moment. You can still keep working the raps and screaming if you want to, right. Some human nature, but you can actually solve your problem by doing these two quick things. The same thing with the feeds, right? They really make very clear to you if you use them and you toggle back and forth between different feeds. Here's my news feed, here's my science feed, here's my friend's feed, et cetera, et cetera, my gardening feed, my cat pictures feed. There's all these different types of communities and feeds that you can subscribe to. And when you see it in that way, you can really see how your social media experience changes depending on what feed you're viewing. And it creates much more awareness, I think, of how your mood is really shaped in part by are you consuming the deluge of doom and gloom and politics and political negativity, or do you have these other feeds available to you that are maybe less toxic or less focused on it's really real.
- [29:08]
A
I mean, there are times when I have my feed up and all I want is just pictures of dogs running around. And I don't want any gloom and doom. And it does change your mood. I want to ask you about two more things before we go. One I think is so important, and I was introduced to this a couple of years ago when I got a briefing, a couple briefings on what AI was already doing and what it could do in the future. And specifically, deep fakes. These are images that really for listeners. I mean, they look and sound real and they're not real. And, you know, I think in the past we could sort of spot what was real and what wasn't. But we're rapidly getting to the point where we're not going to be able to do that. And I was wanting to ask you, how do we, I mean, spot deep fakes? And, you know, I'll give you one example before you answer. I was on a reputable news site recently, and at the bottom of the page, there were some ads that appeared. They appear to be news stories. And, you know, you lead, you go down there and you look at it, and it's a CNN video in which Anderson Cooper is describing how Sanjay Gupta had discovered a cure for Alzheimer's, I believe, and the whole thing was fake. You know, how do we deal with this?
- [30:50]
B
I mean, you know, been writing about this for a long time now, and there's no good answer, really, unfortunately. So there's a couple things that are happening. So first, what you're describing in the video realm has gotten immensely more sophisticated. Even in the last two months, there's been a rollout by OpenAI, something called Sora, right, Where you can really see an entire. It's almost framed as a social network. Like, you can put your face in there, you can put yourself in these videos. Why you would want to, I don't know. People find it entertaining. Again, this question of, like, norms and things shifting is. Is. Is this something that people decide that they like to do for escapism, which is to make animated videos of them, you know, or faked videos of themselves. The problem, as you note, though, is that you can make them of anyone. And I think this question of with public figures in particular, how to treat that is a real open question because there are free speech rules around your right to parody someone, your right to. To do, you know, to make content mocking or criticizing or a political figure that generally has not assumed that you also have the capacity to wholly impersonate them and that that parody has to be disclosed in some way. You know, oftentimes you'll see when somebody would make parody videos in the olden days, you know, with much more challenging editing tools that would require cobble things together, there would be some kind of disclosure that would often go along with it. Whereas now, because of the prevalence and the ease and the relative effortlessness of doing this, it means that it really has enhanced the ability for bad actors to do it and to just push things out and flood the zone where the risks are. Well, bad actors.
- [32:52]
A
And we've seen retweets from the highest office in the land.
- [32:57]
B
Right.
- [32:58]
A
Of fake stuff.
- [33:00]
B
Well, again, I mean, I think in this particular case, people who consider it politically advantageous and choose to ignore the sort of morals of why they say after the fact, oh, everybody knew that was fake. And you can read in the comments that that's just not true. And that's something that we see a lot of. Also, people are being tricked and deceived by these things, which has led to platforms saying, okay, what should we do about it? Right? Because again, you want to preserve free expression, you want to preserve parody and satire. The question is, are you enabling the production of content that is inherently deceitful and how do you do that? How do you make people aware of it? So this question has been, who does the labeling? If the bad actor isn't going to do the labeling, is it the platform's job to do the labeling? And the platforms have looked at this question also because they are carriers of the content, legally, they aren't liable. Legally they don't have to do anything. Right. But being kind of good digital citizens, you have seen Google and Meta try to figure out, should they throw labels up when they can detect something? Is AI generated? I think generally speaking the answer to that is still yes. We should have these disclosure labels. And if the platforms can detect, the platform should do that detection. But there's a lot of nuance to it that makes it hard to just say like labeling is the answer. So first, if you edit an image in Adobe Photoshop and it has AI editing tools now and you describe your edit and you have, I'll use a real example, a picture of Mount Fuji. The photographer used AI editing tools to remove some kind of dust and lens flare, sun flares or whatever they call them. And that picture was uploaded to a meta property and the metaproperty attached a label that said imagined with AI. And so in that imagined with AI was what they said their labels were going to say. It might have just said made with AI. In this particular case, the challenge for that photographer is that's a real picture that he took of Mount Fuji. He just used some post production edits to make it nice, to make it better. Was it generated with AI? No. Is it deceptive? No. Is it useful to label something that is neither? You know, that isn't really deceptive in this way. Well, that's kind of a, you know, you don't want people to get label fatigue either where they just start scrolling past. So that question of when do you apply these labels, who is responsible for it? You know, in the best cases you're going to have open source models and not the best cases, sorry. And the worst I guess is the appropriate term. But in some cases, you're going to have open source models where it's not going to be necessarily so easily detected. These models are not going to proactively produce the watermarks the way that Sora and some of the closed source models proactively produce watermarks on their content. So this question of in the open source realm, what is the open source model community going to do to try to solve this problem? Because it's not only politicians and famous people whose faces are being used. It's also oftentimes women whose images are used in adult content, revenge porn. People who are targeted for retaliation by their neighbor who was mad at them, a local community member who was mad at them. And when you can make a video of somebody saying something horrible, that's where you start to see the question about anxiety come really directly into it.
- [36:50]
A
And we got to have some smart, common sense people that can look at this and have come up with some laws, with tech obviously in the room, because, I mean, it just seems like it's the wild, wild west and it's getting worse and it's going to hurt people. But I do want to ask you and one final really important question. Your work, Renee, seems to be pretty difficult to stomach sometimes. Just all of that you see. And you've been a truth teller throughout this whole process. You've personally faced lawsuits. You talked about some of the harassment that you have encountered, and I imagine it's brought some emotional toll. What, what keeps you motivated to keep doing this work?
- [37:48]
B
Well, I think it's really important. I mean, you don't get targeted if you're not doing good work. So amen to that. You know, let them be mad. Right? Well, the. Now, the work that we got targeted for was work documenting lies in the 2020 election, right, by the President of the United States. I think that's really important work. I only wish that the university had treated it as important work and continued it, but it didn't, and it made its decision. However, I think that that work needs to continue. I think that American researchers need to be studying American elections, particularly the next one, because normalizing lies, normalizing rumors, saying that, oh, we're just going to throw up our hands and say anything goes by reframing fact checking as censorship, labeling as censorship, downranking as censorship. Everything is censorship. Everything that moderately inconveniences somebody with political power is censorship. No, no, no, no. It's the opposite, actually. It's the people who have power, who want to maintain power, who are setting that frame that's why it needs to be challenged, actually. And that's why they work needs to keep happening. Because fact checking, powerful people, labeling the content of powerful people enables ordinary people to know what is actually happening, to know what is actually true, or even just to be made aware that something is disputed. And they should go look for more information somewhere else, which is what the majority of the labels that social media platforms put up on content says. It's just to let you know that, hey, this content is disputed and you should go a lot, you should go look elsewhere. I think that it's really important that that kind of work continues because we cannot have a democracy if we can't agree on the most basic facts. If we can't agree on the fact that, for example, Joe Biden won the 2020 election indisputably, that is true, right? And the democracies depend on peaceful transfers of power. They depend on the public understanding what has happened in an election, what has happened in a campaign, what has happened in the the time and the record leading up to that. More broadly than elections and voting, or maybe I should say more, less partisan than elections and voting, you have collective action problems where society has to come together to make decisions on things. Covid early Covid is a great example. There are going to be a million different stories and rumors and unclear facts floating around. And it's not that you want any of those facts to be taken out of the conversation, but you do want to be able to say, these appear to be more reputable. These are the ones that you should take more seriously. This fake cure over here is probably only going to hurt you. So we should label that and we should let people know that this is not really a thing that they should be taking as seriously as this thing over here for their own personal health. And also when we talk about collective action, as we think about what policies should look like to respond to collective threats, be that climate, weather, upcoming hurricane, all of these things. When we have to make a collective determination, we need to be working with the same set of facts. And that's why I think this work is so important and I plan to continue to do it.
- [40:56]
A
Well, your work is very important and for listeners. Go out and get this book Invisible, the people who Turn lies into reality. Renee, thank you so much, so much. We appreciate you coming on this show. Thank you for being part of the Truth in the Barrel and for everyone. Thank you for watching, listening. Make sure you like and subscribe to the channel. Make sure you follow all of our socials and stay updated. You've listened to Renee as to how to make sure that your social media feeds are a little bit more truthful, shall we say? So thank you. Until next time.
- [41:30]
B
Cheers.
- [41:30]
A
And Renee, thanks for being a part of this.
- [41:32]
B
Thanks for having me.
- [41:38]
C
Morning Zoe. Got donuts.
- [41:40]
D
Jeff Bridges, why are you still living above our garage?
- [41:43]
C
Well, I dig the mattress and I want to be in a T Mobile commercial like you. Teach me.
- [41:49]
D
So Dana oh no, I'm not really prepared. I couldn't possibly at T Mobile get the new iPhone 17 Pro on them. It's designed to be the most powerful iPhone yet and has the ultimate Pro camera system.
- [42:02]
C
Wow, impressive. Let me try. T Mobile is the best place to get iPhone 17 Pro because they've got the best network.
- [42:09]
B
Nice.
- [42:10]
D
Jeffrey, you heard them.
- [42:11]
C
T Mobile is the best place to get the new iPhone 17 Pro on us with eligible traded in any condition. So what are we having for launch?
- [42:20]
D
Dude, my work here is done.
- [42:22]
E
The 24 month EBITDA on experience beyond for well qualified customers plus tax and $35 device connection charge credit send and balance due if you pay off earlier, Cancel Finance Agreement. IPhone 17 Pro 256 gigs $1099.99 and new line minimum $100 plus a month plan with auto pay plus taxes and fees required. Best mobile network in the US based on analysis by OV speed test intelligence data 1H 2025 visit t mobile.com Fall.
- [42:39]
F
Is all about cozy comforts and with HeroBred you can enjoy all your favorites and still hit your health goals. From breakfast bagels and meal prepped enchiladas to mouth watering burgers and cheesy noodles you won't believe. Herobred's options have 0 to 5 grams net carbs and are high fiber from the taste and texture and right now Herobred is offering 10% off your order. Go to Hero Co and use code fall25 at checkout. That's fall25 at HeroCo. All figures are per serving of HeroBread contains 2 to 18 grams of fat per serving. See the product Nutrition Panels on Hero.co for more information.