Transcript
Speaker: Thanks for tuning into CARP Radio, the official podcast of George Mason University's Character Assassination and Reputation Politics Research Lab.
Speaker: Each episode, we bring on new guests to explore the science and theory of reputation, character assassination, and scandal in politics, entertainment, and science, as well as how individuals and organizations can protect themselves against attacks.
Speaker: My name is Yerjane Prega, and I'm a graduate student at George Mason University and the media director for the CARP Lab.
Speaker: For updates on our work, including information about our summer conference on character assassination, illiberalism, and the erosion of civic rights, check out our website at carpresearchlab.org and Twitter at carp underscore lab.
Speaker: Today I'm really excited to welcome my amazing guest, Dr. Josh Braun.
Speaker: He's an associate professor of journalism at the University of Massachusetts Amherst, whose work examines sociological questions surrounding online media distribution.
Speaker: He has been published in many major scholarly journals and his book on the digital distribution of TV news, This Program is Brought to You By, was published by Yale University Press in 2015.
Speaker: Josh is the co-editor of Distribution Matters, a series of books on media distribution from the MIT Press, as well as a founding member of the Culture Digitally National Sciences Foundation Working Group.
Speaker: Josh is also a former science journalist who worked as a contributor to WNYC's Radiolab.
Speaker: In recent years, he has begun a line of research into the ways problems with the digital advertising ecosystem affect the health of news media, and he is currently working on a new book for MIT Press about the civic impacts of media distribution.
Speaker: Josh, I'm so happy to get to speak to you today.
Speaker: Happy New Year.
Speaker: I hope you got a good break from work because when I was reading your preprint, it was pretty dense stuff.
Speaker: I had to engage a lot of brain power.
Speaker: So I'm sure I hope you got at least a little bit of time to step away from that.
Speaker: Well, thanks so much.
Speaker: And Happy New Year to you as well.
Speaker: So I'm really excited to pick your brain about your research.
Speaker: But first, I really wanted to ask you about your career trajectory because you have a really interesting resume and you have a master's and a PhD in communication from Cornell, as well as a master's in bioethics from University of Pennsylvania.
Speaker: And you have a background as a science journalist.
Speaker: So I was really curious about what got you to this point in your career
Speaker: And kind of what came first?
Speaker: Were you interested in journalism?
Speaker: Were you interested in science?
Speaker: And what took you to this point?
Speaker: Yeah, so I mean, I started out, I guess maybe the place to begin is just as an undergraduate, I was interested in science journalism, basically.
Speaker: I started out thinking I wanted to be a documentary filmmaker or something along those lines.
Speaker: But I quickly realized that neither science nor journalism was something I really wanted to give up at the time.
Speaker: So I was able to do kind of create like an individual concentration, basically design my own major with the help of professors at my undergraduate institution, UC Santa Barbara.
Speaker: And as I was finishing up, I worked for a little while in Los Angeles, like picking up odd jobs and like it.
Speaker: But eventually decided I wanted to go back to graduate school.
Speaker: And I was still at the idea that I was going to continue to pursue science journalism.
Speaker: And at that point, you kind of have a couple options.
Speaker: You can go and get a master's in journalism or something along those lines.
Speaker: or you can pursue something that's a little bit more topic specific that might help you with what you want to cover.
Speaker: And so medical ethics seemed like something that would be really interesting to do.
Speaker: And from the work that I'd done just as an undergraduate thesis and that type of thing, I had a little bit of exposure to that topic and the program at UPenn.
Speaker: So I headed over there.
Speaker: And then
Speaker: Um, afterward, I, I took a year off to like, I took a year to actually working as a, as a science journalist.
Speaker: Um, as you mentioned, I like, I, I briefly contributed, um, to like, as a freelancer to WNYC.
Speaker: Um, although only one of those stories ever saw the light of day.
Speaker: So I shouldn't, I don't want to inflate my resume here.
Speaker: Um, and the, uh,
Speaker: But my day-to-day job was working as a junior editor at Seed Magazine, which was sort of its aspirations were that it was a startup magazine and its aspirations were to be something like wired, but more focused on science and a little bit less on technology.
Speaker: And so I worked there for a year, but I was still interested in continuing to pursue graduate education at that point.
Speaker: I'd had a good experience with an advisor at
Speaker: at UPenn, Charles Bosk, who first introduced me to some of the topics that are covered in the paper you mentioned.
Speaker: And so I went at that point, I went and did my PhD at Cornell University in communication, which involved picking up
Speaker: that other master's degree along the way.
Speaker: Um, and the, the big kind of shift that happened while I was there is, um, you know, I'd had a lot of exposure to, uh, this kind of subfield, uh, science and technology studies over the years, um, thinking about kind of how, uh, how science is a social project, um, how technology is culturally inflected.
Speaker: Um, and, uh, and, and at the time there was kind of a shift going on in, uh, in communication research for a lot of people, um,
Speaker: you know, they were grappling with the rapidly sort of changing landscape that the internet was, was ushering in, um, or people's response to the internet was, was ushering in, you know, the media landscape was changing very rapidly.
Speaker: Um, and people were beginning to take the, or to borrow these tools from science and technology studies and kind of turn them back onto the media itself and ask questions about how media technologies, um, were sort of changing, um,
Speaker: the way that the way that the media landscape worked and how they were also and what kind of critical questions could we ask about technologies in our own field.
Speaker: And so that's sort of how I came to what I do today.
Speaker: I study, you know, I study things at the intersection of kind of media technology and culture.
Speaker: And I just I became really interested in sort of in how media is distributed.
Speaker: So, you know, the fact that like so much information is available to us so instantaneously, and yet some things are still really weirdly hard to get because of various arrangements of like digital rights management or copyright or all sorts of different sort of cultural and technological questions.
Speaker: Those those things ultimately became the things that really interested me in the longer term.
Speaker: And that's kind of what I've been doing ever since.
Speaker: Personally, this area is so interesting to me, analyzing the systems of distribution,
Speaker: and applying that to media.
Speaker: And I think it's maybe coming a little bit more to the forefront in the public consciousness as well with the new open AI tools, chat GPT and these other AI tools and how it's led to a new ethical discussion.
Speaker: I was also thinking about one of the topics in the paper we both alluded to, you were discussing how
Speaker: big tech companies try to create better moderation tools and more efficient systems that are more resilient to the type of problems that can create disasters in the digital world.
Speaker: And one of the techniques has been to use AI tools as a form of moderation.
Speaker: There's been a lot of critique of how that works, how it doesn't work, and how algorithmic bias and other issues might be exacerbated by the use of algorithms to try to stem disinformation, misinformation.
Speaker: So I was wondering if you could talk a little bit about that and how companies are trying to harness algorithms to deal with the misinformation, disinformation problem.
Speaker: Have there been successes?
Speaker: Have there been failures?
Speaker: Are there any ways that you see potential benefits coming out of this technology?
Speaker: Well, I mean, so it's a great question.
Speaker: Ultimately, I think moderation is something that's almost always going to have to be done by humans to some extent because so much of it is rooted in heavy cultural context.
Speaker: It's really difficult to understand things like satire or
Speaker: Or, you know, especially when it is like very enmeshed in like particular cultures, subcultures, linguistic norms, you know, all these different things.
Speaker: And so, you know, the person who comes to mind as being like an amazing scholar on this is Sarah T. Roberts, who's written a lot about sort of about content moderation and how it's outsourced frequently to developing economies and things like that, where you have workers who are doing this work.
Speaker: you know, constantly.
Speaker: And there are these sort of cultural barriers to like that are frequently or there are frequently there are frequently issues of sort of
Speaker: I like sort of cultural literacy across cultures that come into these things.
Speaker: And, you know, you can quickly see how that's going to be all the more difficult for any kind of automated system, because, you know, the very nature of kind of online social discussion is we're always talking about new things, new events, new cultural trends, stuff like that.
Speaker: And
Speaker: And so there's always new context to be understood and always new inside jokes and new kind of like forms of irony and that are going to be continually evolving.
Speaker: So I think it's really difficult to suggest that like that AI is ever really going to be able to solve those kinds of problems since it's just a constantly moving target.
Speaker: At the same time, I think most people who study content moderation and just the really traumatic aspects of it, I mean, content moderators who've seen more beheadings than they've been encountered, the content moderators who have seen...
Speaker: You know, child sexual assault in like and basically it's just it's a really just soul rending job.
Speaker: And so I think, you know, on some level, there is a lot of openness to the idea that, you know, if we can automate some of this, that no person should have to do those kinds of that kind of work.
Speaker: So I think definitely there's going to be there are going to be attempts to use more and more artificial intelligence to do this kind of thing, kind of thing.
Speaker: I do think that there are going to be some sharp limits to sort of what those systems can accomplish.
Speaker: And I think, you know, the people who've looked at this, like like Sarah Roberts, like most thoughtfully, I think have pointed out that.
Speaker: One thing about this is just, I mean, just the way that these services, the big issue is often not the technology, but the business model behind it, right?
Speaker: That, you know, these companies...
Speaker: work by creating or they're they they sort of uh their their their gambit is to basically um just is to capture massive scale like to get as many people using their platforms as possible um to get content and to get like this you know an enormous amount of content um and basically at that scale uh you know like those moderation problems become like uniquely
Speaker: in like intractable or really difficult to keep up with.
Speaker: And if you had smaller scale systems, I mean, the argument goes, or a system or systems where like, you know, people were forced to post more slowly or some kinds of like contemporary contemporaneous, like where live streaming was a privilege that not anybody could do
Speaker: within a few seconds.
Speaker: And I obviously realize that there are implications to that that require their own set of ethical discussion.
Speaker: But basically, I mean, the idea that anybody can go live, obviously, with video, which is one of the hardest kind of forms of things, one of the hardest kinds of content to moderate, should we
Speaker: force users to build up some level of trust on these systems before they can do something like that.
Speaker: There's all these different questions about basically the scale at which these systems operate.
Speaker: And maybe we would be better served by sort of systems that didn't operate at that scale or that had sort of a more manageable ratio of human moderators to users or a more manageable like sort of flow of posts, you know, and
Speaker: or a different relationship between moderators and the content they're assessing.
Speaker: So there's all sorts of different ways of thinking about that problem that people like Sarah Roberts or Tarleton Gillespie have kind of dug into.
Speaker: And yeah, I think that the ways forward
Speaker: You know, AI systems are going to probably play a role, but I think it's also important to appreciate like kind of other aspects of these systems that might need fixing at the same time or in place of more automation.
Speaker: Right now, it seems like we're stuck in this really paradoxical place where so much early tech work was about trying to make the experience of using platforms very frictionless and easy and enabling more use.
Speaker: And now companies are circling back like, hey, that was a problem.
Speaker: Maybe we shouldn't have done that.
Speaker: Maybe we need to insert more friction into our platforms and make them more difficult and annoying to use.
Speaker: I don't know how that's working because the first thing that comes to mind to me is on Twitter when people try to tweet something like, you're the most horrible person in the world.
Speaker: I want you to die a thousand times kind of thing.
Speaker: And that little pop up comes up.
Speaker: Most people don't post me in things like this.
Speaker: Are you sure you want to post that?
Speaker: And it's one of those questions like, I'd be really interested.
Speaker: How many decisions to post?
Speaker: Has that changed?
Speaker: So, yeah, it's really difficult to put the chicken back in the egg, I think.
Speaker: I mean, I'd be interested.
Speaker: I would love to hear kind of your thoughts on that.
Speaker: I mean, I think you've probably spent more time considering those kinds of systems than I have.
Speaker: So I'd love to hear what you have to think.
Speaker: What do you think about that?
Speaker: I honestly am very skeptical that it does.
Speaker: I think a lot of people just shrug off that type of stuff.
Speaker: And I wish that there was a way to actually generate more reflection about people's online behavior, but I'm very skeptical of it.
Speaker: I think I kind of keep returning personally to the idea that it's not always just the structure of the systems, but the whole technological medium itself.
Speaker: I think it just does something to our brains that is not good in so many ways.
Speaker: And I think there's too much immediacy.
Speaker: I do agree with the idea that you need more friction, but I don't think you can put friction in the form of pop-ups.
Speaker: I think it needs to come from the actual devices themselves.
Speaker: because the devices enable you to consume so much in a certain way and I don't think the phone that was hard to use would sell that many SKUs after people have gotten used to having these very advanced fast tablets, iPhones, etc.
Speaker: but you know maybe we'd all be better with flip phones again and Blackberries who's to say
Speaker: Yeah, I mean, it's a great set of questions.
Speaker: I mean, you know, there's been the recent, I've spent a lot of time on Mastodon recently, you know, which has kind of emerged as one of the potential alternatives to Twitter.
Speaker: And a lot of the sort of discussion is about like sort of useful forms of friction.
Speaker: And, you know, it's clear that there's not been any, you know, people are still grappling there with like sort of
Speaker: how to prevent and tamp down different forms of abuse and harassment.
Speaker: Um, and there's things that seem to work better about the system there and things that seem to, and seem things that, that, uh, like that don't seem all that effective.
Speaker: Um, but you know, like a lot of the conversation is basically about that kind of, that, that, that, uh,
Speaker: those kinds of questions is sort of like, what is useful friction?
Speaker: Where does it need to be in the system?
Speaker: And kind of, yeah.
Speaker: And so those are really important conversations to have for sure.
Speaker: I don't know why I jumped into this.
Speaker: I think your first answer had so many great points to jump off of that I just for some reason jumped into this one, which is probably not the most useful thing to start our conversation off with.
Speaker: So maybe I just want to go back a few steps and go with a question that gets into your work better.
Speaker: No worries.
Speaker: A lot of your work is on misinformation and disinformation and how that's traveling through these digital channels.
Speaker: The disinformation topic is really highly politicized internationally.
Speaker: And even the definition of what is disinformation and what's misinformation is a battleground.
Speaker: So I wanted to ask you, how do you define disinformation and misinformation?
Speaker: Do you draw any lines between them?
Speaker: And how do you conceptualize that?
Speaker: Well, that's a great question.
Speaker: I mean, I've kind of...
Speaker: My like my my work is kind of focused a little bit more narrowly around basically around
Speaker: you know, the what initially was termed fake news, which quickly became a problem, quickly became like a term that was rapidly problematized and like and politicized.
Speaker: But these these kind of, for example, hoax news sites that would pop up with headlines, Pope endorses Trump, Hillary war and earpiece in the debate.
Speaker: Now there's various like versions of that for covid disinformation and so on.
Speaker: And so I think like there are lots of really important definitional questions around it.
Speaker: I think, I mean, my questions have largely been about how people kind of
Speaker: how people are attempting to monetize different forms of problematic information, whether it's hate speech or hoaxes or disinformation of other sorts.
Speaker: I've focused on the business model around that kind of thing.
Speaker: And it paid a little bit less attention to like to drawing strict definitional boundaries around it.
Speaker: But I think those are really valuable conversations to have.
Speaker: And I do like sort of I do try and track the people who are who are like who are like trying to draw those boundaries like Claire Wardle or Carolyn Jack and so on.
Speaker: And I think they've done a lot of really useful work about kind of exploring both like sort of what do like, you know, what do terms like propaganda or or compromise or whatever else you want to or, you know, what are these terms that like kind of intersect with these fields in different ways?
Speaker: What are the different implications or histories they carry with them?
Speaker: And also, like, how might we parse the like the like lines between disinformation or misinformation or
Speaker: like malinformation, which might be like, you know, it's Claire Wardle and Derek kind of put it like kind of the, you know, information that might not be untrue, but is like sort of decontextualized or put out there in a context where it's likely to do more harm than good.
Speaker: And so there's, yeah, so I think like there are really important definitional questions around that, but, you know, they're not ones that I've like, that have been front and center, I guess, in what I've been looking at for the most part.
Speaker: That's one of the things I really appreciated about your work is how it basically shows that regardless of the motive behind putting out this information, it's spreading through the same channels and through the same weaknesses that are baked into the digital ecosystem, as well as that I think there's this popular idea
Speaker: You know, stories are all being planted in astroturf by political factions or foreign governments, but a lot of it is by people who probably have no other motive than to make money online fast and take advantage of what they can.
Speaker: So I wanted to ask you to talk more about that and go into the channels by which disinformation can spread online by actors who are primarily interested in making money.
Speaker: Yeah, I mean, it's this was something that was you'll recall that like in the wake of the 2016 election, there were a bunch of stories and exposés that started coming out about people who ran these sites that were posing as local news organizations or other kinds of news sites.
Speaker: But they were running stories that were ultimately, you know, hoaxes, Pope endorses Trump, all that kind of stuff.
Speaker: And
Speaker: The reporters had tracked down the proprietors of these sites, and some of them were teenagers in Macedonia.
Speaker: Some of them were broke college grads.
Speaker: Some of them were former reporters, but they were making tens of thousands of dollars off of these hoaxes that they were posting.
Speaker: through sort of ad tech, basically.
Speaker: They were using things like, they were using Google services and others to sort of, their systems for placing automated ads on those sites and they were raking in cash based on the page views they got.
Speaker: And so there were a lot of questions that a lot of people had in the wake of that reporting.
Speaker: And some people were, there were breathless stories about, did this kind of misinformation or disinformation, whatever you'd like to call it, sway the election.
Speaker: There were people who were asking, why do people read this stuff or why do they forward it?
Speaker: What are their mechanisms for likes or what are their media literacy mechanisms that they're using to parse whether this stuff is true or false?
Speaker: And, you know, I'm always fascinated with infrastructure questions.
Speaker: And I was kind of more interested in the fact, like, what is the system that's creating these financial incentives?
Speaker: How is it that you can make...
Speaker: $30,000 off of, you know, off of a hoax news story about ballot stuffing or something like that.
Speaker: And I wanted to see like sort of who's building those systems, you know, how did they think about their work, about their sort of moderation responsibilities or those sorts of things.
Speaker: And so and that was sort of the those are the that was the set of questions that got me interested in this to begin with.
Speaker: And I think one thing that's always important to point out is when we talk about these sums that seem kind of impressive, like $30,000 for a hoax story or something.
Speaker: I mean, those are amounts of money that are wholly insufficient to fund an actual local news organization, but it's a great payday for the broke college graduate with student loans to pay off, right?
Speaker: And so it's that sort of kind of imbalance that I think is a really kind of important aspect of all this.
Speaker: And yeah, so I mean, there's basically this system that exists within the digital advertising world, sometimes collectively referred to as ad tech.
Speaker: It's a system for kind of placing automated ads or placing ads against content in an automated fashion.
Speaker: And there are lots of people who are ready to take advantage of that to sort of to run ads against anything that they can.
Speaker: anything that they can get people to click on.
Speaker: But the last thing I'll kind of say is that it can also give the impression that there's tons of traffic to these sites.
Speaker: And obviously, if there is a ton of traffic to some hoax news story or something like that, it's a good payday for the proprietor who's putting that stuff up there.
Speaker: But oftentimes, a lot of the money involved actually comes from like
Speaker: from fraudulent sources.
Speaker: Like basically they will buy traffic from bots or click workers.
Speaker: And if they can sell it to advertisers at a higher rate than they purchased it for, it's a nice little arbitrage scheme.
Speaker: But there are systems out there for detecting that kind of automated traffic.
Speaker: And so what you need if you're running one of these sites is enough human traffic to confuse those detection systems.
Speaker: And so oftentimes like the, whatever human traffic, one of these like hoax stories can get
Speaker: through its virality, the main purpose of that may be to just serve as cover for the bots that are kind of visiting the page more regularly and drawing the more regular paycheck.
Speaker: So yeah, so the whole thing is convoluted and morbidly fascinating.
Speaker: And it's certainly become a business for people who are engaged in these kinds of things.
Speaker: I was aware of the way that bot farms are used to fraudulently inflate clicks and views, but I was definitely not aware of what you study in ad tech or digital advertising, real-time bidding, and how programmatic advertising actually functions behind the scenes.
Speaker: So for people who are listening like me who have never heard of programmatic advertising, can you explain what it is and how those networks work?
Speaker: Obviously, they're very complicated, so we can't explain everything you know, but just a little overview.
Speaker: Yeah.
Speaker: So, I mean, it's, um, basically when you load a webpage and you see an ad, assuming you don't have an ad blocker turned on or something like that is an increasing number of us do.
Speaker: Um, you know, the ad that you, you are seeing has usually been sold in like just milliseconds, uh, like, like just as you're loading the page through an automated auction.
Speaker: And so it like, so, uh, programmatic advertising is a system basically for automating the, like the, the buying and selling of online ads that appear on web pages and mobile apps and increasingly on like, uh, you know, like, uh, smart TVs and things like that.
Speaker: Um, and, and so there's a digital auction block basically that's called the ad exchange.
Speaker: Um, and,
Speaker: The ad exchange is basically where these, like the, this, where, like, so let me back up for a second, because you're right, it is kind of like, it is a little bit of a complicated topic.
Speaker: So, okay.
Speaker: So when you load a web page and you see the ad, basically, as you load that page, there is an auction going on for your attention, basically, in many contexts.
Speaker: And so the publisher has space reserved for that ad.
Speaker: And they have collected a bit of information about you, the user, with the help of different ad tech services.
Speaker: And they are basically connecting to – and they basically put your attention up for auction.
Speaker: They say, we've got this person here who's about to view our page.
Speaker: And, you know, and here's what we know about them or or here's enough information that you, the advertiser, might be able to call up some memory of what you know about them.
Speaker: And would you like to bid on their attention?
Speaker: Like, how would you like to bid to place an ad in front of that person right now?
Speaker: And the advertisers on their end will bid on your attention, basically.
Speaker: They'll bid to place an ad in that space based on what they know about you and like sort of how desirable they think you are as a consumer, how likely they think you are to make a purchase at that moment, which of course is like in what all those judgments are based on the vast amount of data that people there is being collected about us constantly online and increasingly in real spaces, you know, our phones are kind of collecting location data and so on.
Speaker: So the publisher and the advertiser are both connecting to this digital auction block where this automated ad, this automated auction is going on.
Speaker: That's called the ad exchange.
Speaker: But of course, to run an auction that quickly, because it takes the page just a few milliseconds to load.
Speaker: And by the time it's loaded and you're seeing the ad, all that auction has to be over and the winning advertisers like sort of ad has to be like, is the one that ultimately loads and you see it on the page.
Speaker: So all this has to happen instantaneously.
Speaker: which means that all of the sort of bidding strategies, the demographics that you think you might be interested in, all the things that advertisers might want to sort of use to adjust their bids or decide which publications they might want to run on or which audiences they're interested in, all that stuff has to be programmed in advance.
Speaker: So the advertiser actually uses another service,
Speaker: um which is called a uh sorry um so which is called a demand side platform or a dsp and i'm gonna limit the jargon here because a lot of it um but they connect to basically another service uh to like sort of program in their bidding strategy effectively um and the publisher connects also to the to the the auction block the ad exchange with uh with a um
Speaker: with their own service called a supply side platform.
Speaker: So I'm not going to get any more complicated beyond that.
Speaker: But basically you can see that there's an advertiser, there's a publisher, and they're both connecting to an auction through different services.
Speaker: And so if you're keeping track, there's the advertiser, the publisher, the two services they use to connect to the ad exchange, the ad exchange itself.
Speaker: So that's five different players in this sort of chain of custody here between like sort of where the ad is, where the ad is like displayed and where and who's purchasing it.
Speaker: And there are a litany of other sort of platforms and organizations that can also involve themselves in some way.
Speaker: There are services for like storing and warehousing data.
Speaker: There are services for like sort of scanning publications where an advertiser might place an ad and deciding whether or not they're quote unquote brand safe.
Speaker: Is this place going to be okay to display a Pepsi hat or is it going to be like a shady locale?
Speaker: There's others.
Speaker: There are verification services that make sure that the ad that was sold was actually displayed.
Speaker: There's a ton of these different kind of niche services that are involved in these things.
Speaker: And so there's they're just very there.
Speaker: It's a very complicated sort of chain of custody for the average for the ad to go through.
Speaker: And it creates a lot of opacity.
Speaker: It means that advertisers are often like not they often have no idea ultimately what site the ad they purchase is appearing on.
Speaker: And that's one of the ways in which you get lots of kind of, I guess you'd say less desirable sites, whether it's hoax sites or hate sites or things like that, that are able to make money off of advertising, places that advertisers wouldn't touch necessarily if they knew exactly where their ad was going.
Speaker: But that but that frequently will sort of like be what can be beneficiaries of these sorts of automated advertising systems.
Speaker: So hopefully that wasn't too complicated.
Speaker: I know it's it's kind of a kind of a wild world.
Speaker: One of the things I wanted to address was kind of buried in what you were discussing about how companies try to instantaneously screen whether a website will be brand safe for them to place an ad.
Speaker: However, th is is clearly not foolproof because as you also discussed in your article, normal accidents in the digital age, how programmatic advertising became a disaster.
Speaker: you have a really startling statistic that just from March to July in 2020, in the opening months of COVID, some researchers estimated that websites promoting conspiracy theories and COVID-19 disinformation had collectively raked in about $25 million in advertising revenues.
Speaker: And ads were placed on these websites from major brands like Pepsi, Comcast, and the CDC.
Speaker: You also mentioned that they specifically sometimes favored this content, which is widely viewed as pretty disreputable over other types of content more widely viewed as reputable, like straightforward news and war reporting, activist topics.
Speaker: I was curious, what is it about this type of junky content that makes it more attractive for placing these ads?
Speaker: and leads it to slip through the radar while other websites that promote, for example, war reporting might be considered not brand safe.
Speaker: Yeah.
Speaker: So, I mean, so basically, you know, the way that brand safety is measured is often somewhat, or the way that like sort of the detection services that like are looking for brand safety, like considerations work,
Speaker: I mean, they can have various varying levels of sophistication and you can have like and you can have like sort of an often but oftentimes they'll just scan for keywords and so forth and they'll see something that that that they think advertisers would like to stay away from.
Speaker: And those are the kinds of keywords that frequently crop up on news sites, right?
Speaker: Where like, you know, if like, whether it's like, so things that and so when when content that probably would have been safe for advertisers, you know, or like, or that, or that, you know, is blocked anyway, the sort of industry term is like overblocking.
Speaker: And there's been lots of overblocking associated with like with LGBTQIA topics with, as you mentioned, war reporting with Black Lives Matter protests.
Speaker: Some of these are, you know, some of those decisions might be kind of misbehaving.
Speaker: knee-jerk or reactionary decisions on the part of advertisers not to advertise against some of those topics.
Speaker: But certainly within the world where you would want to advertise against those, there's the
Speaker: Certainly, those are topics that most of us would probably hope to see advertising, supporting reporting on those issues, supporting websites that are discussing those issues.
Speaker: The fact that those things get blocked indiscriminately means that there's a lack of nuance there.
Speaker: I think, you know, in terms of whether or not those same advertisers are ending up on like sort of are ending up spending their money on hoax sites or on like COVID disinformation sites and so on.
Speaker: I think that that's a good question.
Speaker: I mean, I think like on the one hand, you can I can see that like.
Speaker: if you've got, if you decide you don't want to advertise against COVID information, and this was a huge problem during the height of the pandemic, that advertisers didn't want to be associated with like COVID.
Speaker: It wasn't putting people in a buying mood, or at least at the time they feared.
Speaker: I guess we all spent record amounts on like
Speaker: home gyms or whatever it is that people bought during the pandemic.
Speaker: But there was certainly a worry that COVID reporting was not going to put people in a purchasing mood.
Speaker: And so a lot of local news sites, for example, that were trying to cover the pandemic in their area were having a really hard time keeping the lights on.
Speaker: And they were having a hard enough time before, but basically
Speaker: advertisers were pulling their budgets because they didn't have a whole lot of discretionary income at that point to begin with.
Speaker: And they certainly didn't want to be associated with topics that they thought people were going to be running away from.
Speaker: And so, yeah, so I think, so to your, but sorry, so the
Speaker: Basically, overblocking is a huge problem.
Speaker: Overblocking of COVID content was a major problem.
Speaker: At the same time, one of the ways in which people who run disinformation or disreputable sites work is they don't need...
Speaker: um they they they will basically take whatever ads they can get right and so and so i mean i think that's ultimately the the the um like the crux of it is that um is that they're um is that they they'll they'll take whoever's ad is willing is it will appear and certainly like uh and because oftentimes they're um from like lesser known urls and that type of thing and these blacklists or of disreputable sites are constantly evolving
Speaker: you know, sort of a site that crops up quickly enough or there's new sites being developed rapidly enough, not everything's going to be caught in that filter.
Speaker: And so you might know that, you know, a newspaper is running like sort of is running content that you don't want your ads associated with or something like that because they're doing critical, they're doing
Speaker: Or you don't want your ads associated with war reporting or something along those lines, but you might not recognize the sites at the rate at which they crop up or the content might not be scanned quickly enough.
Speaker: So, I mean, at some level, I'm kind of conjecturing here, but I think certainly the hoax and disinformation sites are willing to run whatever ads they can get.
Speaker: They're not discriminating about who their advertisers might be.
Speaker: And the market is large enough, like billions of ads being served hourly across the web, that there's the potential for sites of all sorts to kind of skim a decent amount of profit over time.
Speaker: So yeah, I apologize if that was, yeah.
Speaker: No, it made perfect sense.
Speaker: Like you were saying, I was thinking about when you hear these numbers, oh, these companies, they made $25,000.
Speaker: It seems like a big number, but really it's like a tiny drop in the bucket.
Speaker: And if they can just get enough ads to sneak through and accidentally get served onto their site, essentially because none of these systems are perfect and a lot of them are not very sensitive, well, that can be a big payday for them.
Speaker: And at the same time, that same amount of ads, if that was running on a standard site, maybe which has kind of fallen victim to overblocking, that could absolutely devastate a legitimate organization.
Speaker: So you have this wicked problem, essentially,
Speaker: that makes it very difficult to make ad revenue if you are somebody who might be overblocked.
Speaker: So it's a very difficult situation to be in.
Speaker: And you've put it better than I did.
Speaker: Oh no.
Speaker: Your summary is better than my commentary.
Speaker: So that's, but yes, well put.
Speaker: In your article, you draw on the work of the sociologist Charles Perrow, who has a very interesting theory about how certain disasters stem from factors that are inherent to certain types of systems.
Speaker: I would love to ask you to elaborate more upon that and explain why systemic factors built into the digital world and the ad tech space inherently and inerexibly lead to these issues.
Speaker: Yeah.
Speaker: So, I mean, I, I was prepping a course a few years back where I had a chance to revisit, you know, one of Charles Perrault's essays on what he calls normal accidents.
Speaker: And it's, and I'd been looking at ad tech stuff for a while, this, the real-time bidding, bidding the system of like sort of instantaneous auctions that we were talking about previously.
Speaker: And it struck me that like one was a very good description of the other.
Speaker: And so,
Speaker: Perot, basically, his notion of normal accidents, it comes from, it originates in the 1980s.
Speaker: He decided to take a look at the Three Mile Island nuclear disaster.
Speaker: And he was not the first person to study this, right?
Speaker: There'd been a couple like sort of blue ribbon commissions and so forth that had been tasked to figure out what went wrong in this nuclear disaster.
Speaker: And Perot,
Speaker: those were mostly about sort of finding accountability, placing blame, determining, you know, exactly like sort of what, what failed.
Speaker: And Perot reached what a lot of people thought of as, and it was his, his work was provocative because he reached what a lot of people thought of as kind of a more troubling conclusion, which as you said, is that it's sort of in certain types of systems like accidents or disasters are just kind of inevitable, right?
Speaker: Like you can't, that they're, that they're going to occur at some point.
Speaker: Um,
Speaker: And the two features that he distills that he thinks are inherent to these systems where accidents will always occur or eventually occur is one of them is, as you mentioned, what he called tight coupling, which basically corresponds to just really high levels of automation or where things happen really quickly in a sequence that's been determined in advance.
Speaker: And there's really little opportunity for human intervention as sort of the dominoes fall, sort of.
Speaker: It's really hard to step in and like stop the process once it's begun.
Speaker: So that's a very tightly coupled example of something that's very tightly coupled.
Speaker: Or that's the notion of something that's very tightly coupled.
Speaker: The other thing that he thought he thought a lot about was this notion of interactive complexity.
Speaker: Well, and so the term like to kind of to get rid of the jargon here, like lots of things can be complex, right?
Speaker: Lots of systems can be complex.
Speaker: What what Perot sort of appreciated was that different systems could be complex in different ways.
Speaker: So to go back to the dominoes thing, right, you can imagine standing up like a thousand dominoes and wanted to see them knock all each of them, knock the others over.
Speaker: And so if you set your dominoes up and you tip them over and all of a sudden one of them doesn't fall, right?
Speaker: your design could have been complex and maybe have thousands of dominoes arranged in a really complicated pattern.
Speaker: But the moment one domino falls and it doesn't touch the next one, you can quickly see what went wrong, right?
Speaker: So in Perot's like sort of
Speaker: like equation, this would be like something that's linear complexity.
Speaker: One thing is leading to the next, to the next, to the next, and you can quickly see kind of what's gone wrong.
Speaker: It's easy to identify the problem.
Speaker: But other systems and the ones that he's concerned about are the ones that are what he called interactively complex.
Speaker: And it's a little bit harder to describe, but it's the notion that sort of you have different components or processes in like the system that you're running
Speaker: that interact in ways that are kind of random and unforeseen and really hard to sort of understand like exactly how these two things are going to affect each other or how these many things will affect one another once they're sort of all going on simultaneously.
Speaker: And so what Perot basically said is that you can have organizations that do a really great job at managing tight coupling, like these high levels of automation and these rapid processes.
Speaker: And you can have...
Speaker: organizations that do a really great job at managing interactive complexity, but it's really difficult to have an organization that's good at managing both things simultaneously.
Speaker: And so, and the reason basically that he comes up with is that sort of if you've got a system that's really tightly coupled, you know, that's basically where everything's unfolding very rapidly and each step determines the next one in a sequence that's so quick that you really can't like sort of intervene.
Speaker: Any sort of fail safe that you come up with, any solution to a problem that's going to happen in that system has to sort of swing into action like really rapidly.
Speaker: And so you need a lot of sort of like plans that are determined in advance about what to do.
Speaker: Right.
Speaker: And if you've got if you've got that, you've got a system where it's sort of like, don't think, just follow the procedure.
Speaker: And that's a very top down way of managing an organization or and.
Speaker: Whereas interactive complexity, where you have when a problem occurs, it's between lots of different components that are not it's not easy to understand how it's how they're interacting with one another.
Speaker: That sort of system requires you to empower people on the ground to sort of troubleshoot and investigate and give them a lot of authority to sort things out.
Speaker: And so they're very different management structures that don't easily dovetail well.
Speaker: And you can talk about them in the sense that like as like sort of like sort of human systems management or automate or different forms of automation or whatnot.
Speaker: But it's really difficult to have both things.
Speaker: And digital advertising has both things.
Speaker: It's it's very, you know, so you've got tons of automation, right?
Speaker: Billions of ads are served hourly online by these systems and you have lots and lots of interactive complexity.
Speaker: Right.
Speaker: So where like you have many, many different like sort of digital systems connecting to each other and interacting in ways that are not always not always well foreseen.
Speaker: And so, and basically it just becomes really hard to manage.
Speaker: And so, you know, Perot basically argued when you have a system that has those features in combination, you're going to get accidents.
Speaker: Like you're going to have, like you're going to have inevitable sort of failures.
Speaker: And so, you know, he was studying nuclear accidents, right?
Speaker: So like a mushroom cloud is a really scary thing.
Speaker: When you have, when you have these kinds of failures in digital advertising, you know, the, like they're not,
Speaker: There's no mushroom cloud when, you know, a Volvo commercial is viewed by a bot or a Pepsi ad appears on a pornography site or something like that.
Speaker: But what you do get is you get like these lots of tiny but rapidly compounding failures that add up to $25 million worth.
Speaker: advertising being placed on COVID disinformation sites or $5.8 billion of ads being viewed by bots that otherwise would have gone to legitimate publishers otherwise, or immense amounts of money lost to overblocking, those kinds of things.
Speaker: So you have really tiny normal accidents, but really big impacts once you add them all up.
Speaker: So hopefully that's a reasonable summary, I guess.
Speaker: It makes perfect sense.
Speaker: And I love the quote you included that said, data isn't the new oil, it's the new oily rag, a low-grade waste product that is only valuable when it is piled up in such vast quantities that it poses an existential civilization-ending danger.
Speaker: It's so interesting that the problems that are inherent to this whole industry, like a Coke ad showing up on an obscene website, because they seem so low stakes, paradoxically, that's why they're so dangerous and they pose so much of a threat.
Speaker: Charles Perrault, he argues that we should be willing to abandon the systems that lead to normal accidents if the consequences of those accidents would be unacceptably high.
Speaker: And obviously, that's always going to be a matter of debate.
Speaker: What's unacceptably high?
Speaker: But, you know, if we both sit here today and we say it's unacceptably high, is it even possible to abandon the system we have?
Speaker: And are there any proposals for a new system that could take its place or change enough parts of it that we could essentially get rid of the digital advertising ecosystem as we know it today?
Speaker: Yeah, I mean, those are great questions.
Speaker: And I think those are the big questions that regulators, for example, are asking at the moment.
Speaker: I mean, so we obviously had advertising before digital advertising.
Speaker: Much more of it was contextual advertising, right?
Speaker: So you want to sell your shoe, you put your ad in a running magazine or something like that, right?
Speaker: And so you were using the publication as a proxy for the audience.
Speaker: The big thing that happens with targeted advertising of the sort that we're dealing with, with like real-time bidding and so on,
Speaker: is that basically, you know, it's what Shoshana Zuboff famously called like surveillance capitalism, right?
Speaker: We're looking a lot at sort of, we're basically trying to surveil users, find out all their behavioral identity traits and like sort of,
Speaker: and use those to target ads and advertisers don't care so much what the website is that serves the ad.
Speaker: They just care about reaching the right person, right?
Speaker: And that lack of respect for context is fundamentally, you know, opacity and like complicated explanations about the digital advertising, you know, like ecosystem aside.
Speaker: like that lack of appreciation for context is fundamentally what gets us here, right?
Speaker: That is the moment advertisers are more concerned about like sort of reaching the person they think is likely to buy their soap than about like sort of whether their ad is appearing on, you know, a neo-Nazi website.
Speaker: That's like, that's the trade-off that like ultimately we're talking about.
Speaker: So I mean, you can imagine like a greater return to contextual advertising
Speaker: And there are there has been that in recent years, I think, like, you know, so advertisers have shied away a little bit from from like sort of the real time bidding like section of the digital advertising market.
Speaker: There are forms of programmatic advertising that allow you to like, you know, that that that are that allow you to do business just with reputable publications or you may decide to do business directly with CNN or The New York Times to post ads on their websites and so on.
Speaker: So, I mean, there is more of a return to contextual advertising in some senses.
Speaker: There are folks like policy critics like Johnny Ryan who advocate basically for changing the standards about what kinds of data can be collected about people, right?
Speaker: So that you don't have as much.
Speaker: So like if you're going to target advertising, maybe you limit the types of information that people can collect.
Speaker: or the types of information that are actionable, right?
Speaker: So if your digital advertising protocol only accepts like, or only like sort of understands a more limited number of like sort of identifying traits or something, then there's less incentive to collect all the additional stuff.
Speaker: You can also put companies on the hook in different ways for things like data breaches and stuff like that.
Speaker: So Cory Doctorow, who's the original author of the oily rags quote, one of the things that he's talked about basically is like, okay, let's basically make companies liable for these data breaches in ways that they haven't been before.
Speaker: And if suddenly it costs a company an enormous amount of money, if the Twitter breach happens and everybody's emails get out or something like that, a company that's deciding about collecting that stuff, they might have to go out and buy insurance, basically, and say, okay,
Speaker: And if your insurance rate goes up, your corporate insurance rate goes up based on how much personally identifying information you're collecting, maybe you decide that, okay, we could do 0.05% better revenue, like something like that, by targeting our ads better, by collecting X, Y, or Z amount of data.
Speaker: But that pales in comparison to the amount of additional insurance we're going to have to take out to basically be keeping records of those things.
Speaker: then that maybe changes companies' incentives or relationships to these types of things.
Speaker: So there's lots of questions about, can we limit the amount of data that people collect?
Speaker: Can we get people to return to contextual advertising?
Speaker: And then fundamentally, there are people who are asking questions about the advertising business model more broadly.
Speaker: So folks like Victor Picard,
Speaker: For example, I talk a lot about basically, you know, that advertising, the advertising business model in some ways has been poisonous for journalism all along.
Speaker: And maybe instead of it was, we're reevaluating the relationship between public service media and, and advertising.
Speaker: Maybe it's, maybe we should be asking fewer questions about how do we get like the good old days of contextual advertising back and be asking more questions about sort of what do we replace it with?
Speaker: Whether it's like some kind of public subsidy or,
Speaker: where you take the winners in the digital economy like Comcast or Google, and you have them pay into funds for public service or investigative journalism.
Speaker: You know, so there's a whole spectrum of like different things that people are kind of considering.
Speaker: And I think, you know, it ranges from like sort of, can we reform the digital advertising ecosystem to can we replace elements of it or at least for the element, the parts of our online lives or our information ecosystem that we really care about the most.
Speaker: I think those are really interesting ideas about how we can take the profit motive, which has been incentivizing so many of these terrible consequences and see can we almost reverse engineer it to prevent at least some of these normal accidents from occurring.
Speaker: When you're talking about contextual advertising, I think a good example of how that occurs right now is influencer marketing is a very big business and all sorts of sponsored content has really exploded on the internet.
Speaker: So it's interesting that even as the digital advertising ecosystem is often so devoid of context, this new form of what is essentially contextual advertising has been birthed online and is sustained online.
Speaker: that's a great example.
Speaker: Emily Hunt has a new book called the influencer industry where she talks a lot about this.
Speaker: And, and I think that that's been kind of inspiring for me in the sense that like,
Speaker: one of the sort of big developments, I guess, in the influencer industry, insofar as I understand it through her work, I should say, since I'm not an expert on it myself, but has been basically that brands have decided to work with influencers who control smaller and smaller numbers of followers.
Speaker: It
Speaker: And they have craven reasons for doing so, right?
Speaker: Like, so, you know, as opposed to working with somebody who has like 2 million or 250,000 followers or something along those lines who might be sort of harder to sort of micromanage in the way that the brand would prefer, they can go to basically smaller and smaller influencers, people who might have a few thousand followers or a few hundred followers or something like that, people that they're calling nano influencers and so on.
Speaker: But, and try and work with those individuals.
Speaker: And so, I mean, like I said, while you can question the motives of the companies that are making decisions to do that, it also suggests that there's something analogous to the motivations behind companies.
Speaker: real-time bidding that is being solved in a different space, I guess, in some ways.
Speaker: So when we're talking optimistically about real-time bidding, the thing that it allows, one of the big reasons it exists, is that it allows basically anybody to take advantage of advertising.
Speaker: So you or I can start a blog about our hobbies and
Speaker: or something, and we can immediately start making money from it by placing ads.
Speaker: And that was something that you couldn't do before unless you were a publication that was large enough to command the attention of your regional advertisers or national advertisers, something like that.
Speaker: And suddenly anybody can basically make a little bit of money
Speaker: off of their personal website.
Speaker: And in the same way, like, you know, there if in a world where sort of anyone can be an influencer with a few thousand followers and potentially monetize that through these systems that have been put in place for brands working with influencers.
Speaker: You know, it's it's suggested there like, you know, there is a model for working with small outlets or or or people or people or sites that are not commanding a huge amount of attention, but are commanding like a.
Speaker: smaller amounts of legitimate engagement in a way that is really sensitive to context.
Speaker: And it could be that there are huge problems that I'm missing around the influencer, like sort of these developments in the influencer industry, because that's certainly possible.
Speaker: It's not like sort of, it's by no means my area of expertise, but it is interesting, at least to me, to see that there are other things that are being tried in sort of different areas of sort of marketing and advertising that might provide more of an aspirational model sort of within the advertising space.
Speaker: Something about social marketing and influencer marketing is that in a weird way, it does empower the consumer because the consumer has more of a relationship with those smaller influencers and they can get that like immediate feedback with them, which is not always good.
Speaker: Obviously, we at CARP, we think about character assassination and any viewer or follower can use the same
Speaker: online tools to create a completely unfounded smear campaign against some small influencer and destroy their lives and it's not always great but I was thinking that maybe this could be a way to deal with some of the problems created by over blocking because
Speaker: We were just discussing, you know, we might like to see ads for a company's products next to somebody who is promoting pro-social causes or some sort of valuable work, even if that's not something that advertisers stereotype or will get you into a buying mood.
Speaker: And you've done some work on the group Sleeping Giants, which organizes consumer boycotts, usually in favor of some progressive causes and trying to
Speaker: lobby companies to not place advertisements with certain corporations.
Speaker: And I was wondering, could similar tactics be harnessed to address some of the issues in ad tech or the potential as a consumer, how we can try to use whatever purchasing power we have to address some of these problems?
Speaker: Yeah, so I mean, these are great questions.
Speaker: So I mean, Sleeping Giants and kind of a other campaign, Stop Funding Hate that you might think of as kind of a fellow traveler.
Speaker: you know, one of the things that they, or the thing that they, they sort of did and sleeping giants now has, or at different times has had chapters spread across many different countries.
Speaker: But basically they created social media accounts that, that you could, if you saw an ad in on like placed in, if you see, if you saw an ad on a site,
Speaker: that was distasteful for one reason or another.
Speaker: Breitbart was sort of the prototype here where they were running inflammatory headlines, single people out on the basis of race or gender.
Speaker: And basically people would take snapshots of their sort of most inflammatory headlines, often with like an ad displayed next to them.
Speaker: And they would forward the screenshots to the Sleeping Giants Twitter account, or they would just tweet them out directly and tag the Sleeping Giants Twitter account on social media and basically say, you know, like, hey, Unilever or hey, you know, whatever brand you're talking about.
Speaker: Like your,
Speaker: you know, you have all of these inclusive things that you say in your marketing.
Speaker: And yet here is your headline that's running next to this article that's saying horrible things about migrants, you know, or about LGBTQ people.
Speaker: And so, and Sleeping Giants, which, you know, gathered many, many followers would sort of amplify those tweets.
Speaker: They would also like sort of, if the company responded by saying, oh, we'll address this, we'll take down the ad,
Speaker: They would also sort of like tweet out those follow-up informations.
Speaker: And they would also accumulate a list of brands that had said that they would remove their ads from Breitbart or said that they would remove their ads from these other sites.
Speaker: So they basically created this kind of like sort of document that was intended to kind of hold these brands accountable and also kind of keep a measure of the progress or the sort of success of the campaign.
Speaker: And so...
Speaker: you know, it resembles a boycott in a lot of ways of these brands, like what you, you know, like audience, your audience will probably recognize as sort of like a media boycott.
Speaker: At the same time, you know, the people behind the different Sleeping Giants chapters would say that that's a misunderstanding because they think that these, that these companies were not intending to run their ads on these sites and they were kind of providing public service by like, by allowing them to see that this was happening.
Speaker: And so, and, but,
Speaker: Yeah, I mean, like, they are fundamentally one of the problems that Sleeping Giants or things like Check My Ads, Nandini Jami, who's formerly of Sleeping Giants, and Claire Atkin, sort of her business partner, run the Check My Ads Institute and a great newsletter called Branded that people are interested in this stuff might be interested in.
Speaker: But in any case, they, you know, the things that they're
Speaker: one of the sort of fundamental things that that model kind of sets out to address is that if you take, so if you take like,
Speaker: an optimistic view of brands' motivations.
Speaker: Like, they don't mean to run ads on these sites.
Speaker: And it's happening because these systems are so kind of opaque and they don't necessarily control where all the different ads, and they don't have a great control over where all the ads are going.
Speaker: And so if they can add Breitbart or another site to a blacklist, like, you know, or update or figure out where the glitch was when they attempted to blacklist these sites or something like that, that they will immediately get rid of that.
Speaker: But there is a more sort of... We talked earlier about the fact that targeted advertising privileges...
Speaker: finding the user you want and displaying them an ad irrespective kind of editorial context.
Speaker: And so in that sense, it's kind of like going into a shady bar and seeing somebody, you know, right.
Speaker: Should you be embarrassed or should you recognize, or maybe like, maybe it doesn't matter because they're there too.
Speaker: Right.
Speaker: And like, so at the, and so, and so like, and, and so in the same way, like,
Speaker: advertisers could typically safely assume that, you know, if you were on right part or if you were on another site with like, like a reputation for like things that tow the line with or cross the line into hate speech or other problematic stuff, you know, that user probably intended to go there is the conceit.
Speaker: Right.
Speaker: And so they're not going to complain when they see the Pepsi ad in that spot.
Speaker: Right.
Speaker: Because they don't have a problem with it.
Speaker: They don't have a problem with the content they intended to view that.
Speaker: Well, that's, you know,
Speaker: So it could be that advertisers are acting in a way that's completely amoral and they don't care about their ads displaying these places.
Speaker: Or it could be that they buy hundreds of thousands or millions of ad impressions and when 300 of them show up on Breitbart, that's leakage and they never really cared about it that much because even though they didn't mean for it to appear there, it was such a small fraction of their ad buy that they weren't really concerned about it.
Speaker: So in any case, what Sleeping Giants or Branded with these sort of operations are doing is they're creating sort of reputational risk for these companies where there wasn't any before, right?
Speaker: They're saying, okay, now it's not just between the user who enjoys Breitbart and the brand that their brand is appearing on Breitbart.
Speaker: it's announcing to the world that, you know, that Dove Soap is running ads on Breitbart or something like that.
Speaker: And that may be an apocryphal example.
Speaker: I don't know that Dove Soap ran ads on Breitbart, so I'm not slandering anybody here.
Speaker: But in any case, like that's the, but that's sort of the thing is if you create reputational risk for the brand, like sort of where it changes the proposition of like sort of how safe it is to run ads in that space,
Speaker: And so whether you think that the brand was doing it, you know, like I said, in kind of an amoral way, or if they were doing it, or if they just weren't aware that their stuff was appearing there, but didn't really care, you know, there's all this different like sort of interpretations of what brands motivations might have been or what the incentive structure was around it.
Speaker: But this certainly changes that incentive structure.
Speaker: And so there have been different Sleeping Giants chapters that kind of emulated this that have cropped up in countries around the world with different amounts of longevity or levels of success.
Speaker: There's been a very successful chapter in Brazil recently.
Speaker: Like I said, the Check My Ads Institute, which was spun up when Andy Jami left Australia.
Speaker: And Sleeping Giants has been, I think, hugely successful.
Speaker: And I think these are all good sources of accountability.
Speaker: I don't know that they'll solve the problem for all the reasons that we talked about earlier.
Speaker: Maybe some of these things are kind of intractable.
Speaker: But at the same time, they're really important.
Speaker: Number one, of course, I do not believe you were slandering Dove Soap, Dove Soap, do not sue us, we don't have any money.
Speaker: But it's very interesting as it is a pain that as a consumer, you know, you can't fix any of these problems, even if you can get together.
Speaker: I'm not very familiar with the Sleeping Giants business model.
Speaker: I was aware of them at that surface level.
Speaker: I don't know if they're funded by any larger organizations, but obviously they're a smaller group.
Speaker: Yeah, they're not really a business.
Speaker: I mean, it is just kind of an activist collective.
Speaker: They do sell hats and things like that.
Speaker: But as far as I know, they all go to funding whatever small amount of operational costs that they have.
Speaker: The one thing I would just add is it goes back to the question of whether or not people see advertising as fundamentally kind of a good model for whether it's journalism or other kind of public interest media.
Speaker: Because, you know, sort of ultimately what Sleeping Giants is doing when they're like, you know, when they're calling on these brands to sort of be more accountable with their spending, that could fix a lot of problems with the digital advertising ecosystem.
Speaker: But at the same time, it's also sort of
Speaker: you know, my colleague, Emily West would say like, was one of the coauthors on that, that paper with me was, you know, is pointing out that it's also kind of every time they like an advertiser kind of disciplines a publisher by pulling their ads or something like that.
Speaker: It's sort of, it reestablishes who's in charge of the media ecosystem, right?
Speaker: Like who controls the money and continues to sort of harden that structure of advertising that we have.
Speaker: And there was a time when, you know, progressives like,
Speaker: would have found that like, like basically there was a time when progressives would have found like, you know the idea of like, of turning social responsibility over to brands to be pretty horrifying.
Speaker: And so, I mean, I think it is, we, we do live in a kind of a, like a brave new world in that sense.
Speaker: You know, so, cause I mean the other, if you, if you're less saying when about the role of advertising broadly, you know,
Speaker: You know, you might want a system where like we're not like sort of we're not relying on advertisers to be to be our moral compass.
Speaker: So I just wanted to throw out there's there's more than one side to these campaigns.
Speaker: I do think the work they're doing is really important.
Speaker: I think it also raises like plenty of interesting questions about like sort of the kind of future we want in terms of how our media are funded, particularly the ones like journalism and that we really rely on.
Speaker: Yeah, and it's a question, I think, of working within the system.
Speaker: Is it better to try to use the mechanism of the system to our own ends?
Speaker: Or do we want to totally opt out of that system?
Speaker: And that idea of opting out of the system, we were discussing Shoshana Zuboff's idea about surveillance capitalism and how ad tech is enabled by this massive scale data collection.
Speaker: as consumers and users if we're somebody that would like to opt out of that.
Speaker: Obviously you can opt out of it as you were saying, it's all being tracked all the time.
Speaker: But are there any tools we can utilize to somehow limit the amount of data we're giving up?
Speaker: If someone's listening to this and they're like, I don't like this at all, I know I can't get out of this, but how do I just limit a little bit more all the data I'm serving up?
Speaker: Yeah, I mean, that's a great question.
Speaker: And I mean, a lot of it is kind of complicated by design.
Speaker: I mean, there's a little like sort of triangular logo that you may see on like banner ads and stuff like that, that you can click and you can go and like opt out of the advertisers, like sort of like some of the data collection that those advertisers do.
Speaker: You know, it's, you know, and you can similarly kind of go through your settings in different ways.
Speaker: apps or search engines and kind of like manually delete or opt out of like sort of different levels of tracking.
Speaker: But, you know, I think as Zuboff kind of like sort of underscores, I mean, it's really like hard to do and it puts a lot of the onus on sort of end users to deal with all of these things.
Speaker: I mean, and so
Speaker: Yeah.
Speaker: I mean, so you can certainly use services that track you less.
Speaker: You can, you know, use DuckDuckGo as your search engine or, uh, or, uh, or,
Speaker: Or, yeah, or things.
Speaker: But I mean, I don't have like a perfect or really simple solution for folks.
Speaker: I think one of the one of the things that I talk about sometimes when I discuss media literacy with my students, right, like I teach in a journalism department.
Speaker: You know, we're all about the First Amendment.
Speaker: You know, I am a big fan of freedom of speech and everything else.
Speaker: Yeah.
Speaker: you know, something like the first amendment also, I think it's important to recognize that sort of, if, if you were, if, if you were going to like, um,
Speaker: constitutionally enshrine the ability for anybody to say within, you know, some limits, almost anything like that puts that by default puts a lot of responsibility on like sort of on individuals to have like a really good BS filter, you know, and, and so and, and I think when we think about that in like our contemporary environment,
Speaker: you know, we have these systems, like, I think it's also important to realize that in some ways that is like an individual level, an individual solution to a structural problem, right?
Speaker: In the same way that, you know, auto manufacturers might have suggested that, um,
Speaker: you know, they don't need to install seatbelts.
Speaker: People just need to drive more responsibly.
Speaker: I mean, you know, like there are, you know, and you can name any number of sort of like of analogous sort of examples where, you know, we have a kind of a media ecosystem that,
Speaker: that has a set of incentives.
Speaker: We've been talking about ad tech, but, you know, like other and certainly other other people have talked about plenty of other areas of this, but that basically incentivizes the, you know, creation and spread of clickbait or misinformation or disinformation or creates monetary incentives for hate speech and all these other things.
Speaker: Right.
Speaker: And so expecting individuals basically to compensate for that,
Speaker: But through increased levels of individual responsibility, I think is like on the one hand, you know, it may be something that we need to do, right?
Speaker: For the reasons that, you know, we cherish free speech.
Speaker: So we all need to have like really good BS meters in response.
Speaker: But at the same time, like, you know, I think it's also important to think about like, what are the structural reforms and solutions that don't put all of the onus for having a healthy media ecosystem or sifting enormous amounts of like questionable or ambiguous content that
Speaker: on like individuals.
Speaker: Right.
Speaker: And so, yeah, so I and I don't, you know, obviously what those reforms might be is always going to raise questions about freedom of speech or sort of who has like editorial or like a responsibility.
Speaker: And there's there's lots of really thorny questions around this stuff.
Speaker: But
Speaker: I think, but I always wanna balance like sort of like what can I do as an individual with sort of like, why should we, and I think Zuboff kind of arrives at a similar conclusion, like why should we have to be forced into some of these choices or like have to have these coping mechanisms?
Speaker: Can we live in a system that didn't make us like perpetually defensive or operating from that mindset?
Speaker: And we shouldn't have to.
Speaker: I sometimes feel like it's a little bit like how every academic paper ends with the call for more papers on this topic.
Speaker: That's how I feel like talking about any of this is, you know, it's like, we've diagnosed the problem.
Speaker: We're in need of some more solutions.
Speaker: Yes, yes.
Speaker: And I appreciate that you brought up clickbait.
Speaker: I didn't get to mention it and we've gone pretty long, but I have several questions I had written about skewed incentives and clickbait in journalism.
Speaker: So maybe I'll have to wrangle you back on to talk more about that topic too.
Speaker: Well, it's absolutely a pleasure.
Speaker: Before I let you go, I always like to end episodes with a question about three things or just one thing even that you would recommend people listening to this podcast read or watch to better understand the media environment today.
Speaker: Okay, great.
Speaker: Let's see.
Speaker: So a book that I've really become fond of in recent weeks and months is Kevin Driscoll has a book, The Modem World.
Speaker: And it sort of makes the argument that a lot of the histories of the internet, and by extension, are ways of understanding the internet and what its problems are.
Speaker: have been kind of pretty limited in nature.
Speaker: They talk about like the technical design of it and so on.
Speaker: And he points out that sort of the social precursors to the internet are things like, you know, BBSs and sort of like small hosted spaces where people were congregating online before widespread access to the internet became a thing.
Speaker: And he looks at sort of the history of those kind of small do-it-yourself bootstrapped online communities as being kind of like an important part of online history that led us to the expectations that we initially approached things like social media with and so on.
Speaker: And he also points out that that, if we think about it in that way,
Speaker: it gives us a whole different set of experts and a whole different set of historical perspectives for thinking about the problems with online social spaces now.
Speaker: And so I think a lot of that book and the conclusions that he reaches and some of the questions that he revisits from key moments in that history have resonated really well recently, especially with all of the questions surrounding the future of Twitter and the questionable decisions there.
Speaker: people have kind of begun reevaluating their relationship to like centralized platforms that are not moderated as small communities and so on.
Speaker: So that one's definitely one that would be high on my list.
Speaker: For a lot of the moderation topics that we've been talking about, there's a couple books that I would definitely refer people to.
Speaker: Sarah T. Roberts has a book behind the screen about content moderation that's really extraordinary.
Speaker: And Charlton Gillespie has a book that looks at content moderation questions and asks some really compelling questions about the amount of agency that users have or should have in relation to them.
Speaker: That one's called Custodians of the Internet.
Speaker: Yeah, so I think I can leave it there since I can always suggest more books, but that's three.
Speaker: So yeah, I'll leave it with those.
Speaker: I'm definitely going to be looking at those books because this is such an interesting area where I really don't know anything about this and I want to know more.
Speaker: So I appreciate it a lot.
Speaker: So just thank you so much for taking your time to join me today.
Speaker: I really appreciate it.
Speaker: Well, thank you so much for having me.
Speaker: It's really been a pleasure and it's an honor to be invited.
Speaker: Thank you.
Speaker: I don't want to hold you any longer, but I do want to ask you, if anyone listening wants to find more of your work or read any of your upcoming writings, can you share how they can find more of what you do?
Speaker: Well, see, my university just started offering its own sort of web hosting.
Speaker: So I just shifted my website to there.
Speaker: So I think it's like joshbraun.umasscreate.net or something along those lines, which is where like sort of my professional website is hosted now.
Speaker: And certainly I'll post things that I'm writing there as it comes up.
Speaker: And in terms of where I will be next, this summer, there's going to be a workshop at Cambridge that's being organized around some of these questions around disinformation.
Speaker: And so I have plans to be there.
Speaker: And the details of that have yet to be entirely announced and solidified.
Speaker: But I can certainly add them to the aforementioned website as it comes up.
Speaker: Wonderful.
Speaker: And the link to that will be in the show notes.
Speaker: So I definitely advise everyone to check it out.
Speaker: All right.
Speaker: Well, thank you so much again.
Speaker: It's really been a privilege.
Speaker: Thank you.
Speaker: Thanks for joining us for this episode of the Carp Radio Podcast.
Speaker: If you like what you heard or have any suggestions for future episodes, please leave a comment and a five-star rating on Apple Podcasts.
Speaker: It really helps us.
