Episode 6: Disinformation and its consequences, with Prof. Claire Wardle

SYNOPSIS

Dr. Claire Wardle is a disinformation expert and communications professor at Cornell University. She joins Rachel and Dan to talk about the history of disinformation study, how new technology introduces new challenges, and about how the information ecosystem might change in the future. She offers insights into the psychology of disinformation. From the conversation, Rachel elaborates on the lens of Friction, and Dan talks about a lens he can only call “Middle of the night.” 

STORIES OF DISINFORMATION

The Propaganda of Hearst and Pulitzer

• The Spanish American War and the Yellow Press (Library of Congress)
• The Gilded Age (Wikipedia)

Bedlam and Squalor and Worse, Oh My

• Transcript of August 11 press conference (Roll Call)
• Trump exaggerates DC crime while ordering police take-over and National Guard deployment (Politifact)
• Trump’s DOJ claims credit for falling crime rate in DC (Justice.gov)

INTERVIEW

• Dr. Claire Wardle
• Australian drama series about the wellness industry: Apple Cider Vinegar
• Jim Acosta “interviews” LLM based on Parkland shooting victim
• Dr. Noc, social media science influencer, who may not be scruffy enough to be believed

LENSES

Friction

For designers, "friction" represents how much the experience intrudes on consuming information or completing tasks. Experiences increase the intrusion to discourage destructive actions (like deleting important information). Additionally, it is often said that lies are “low friction” and truth is “high friction”. That is, understanding and internalizing a lie is less cognitive demanding than doing so with the truth. 

• How does the system use friction to engage users’ critical thinking skills? 
• How does the system reduce the friction on high quality information?
• How does the system titrate the amount of friction in the experience relative to the quality of information?

Middle of the night

Users are most susceptible to disinformation when they are at their most vulnerable, like in the middle of the night with a health scare.

• How does the system support users who are agitated and vulnerable?
• How does the format of the information take into account a user who is agitated or vulnerable?
• How would the system fare if it catered to someone using it in an extreme scenario that prio

_____________________________________________________

Personnel

• Dan Brown, Host
• Rachel Price, Host
• Emily Duncan, Editor

Music

• Turtle Up Fool, by Elliot

_____________________________________________________

Unchecked is a production of Curious Squid

Curious Squid is a digital design consulting firm specializing in information architecture, user experience, and product design

Transcript

SPEAKER_01 0:00

The absence for many people of human connection means that even just a connection with a machine is enough. And you see people now very quickly having relationships with these machines because it mimics what they used to have.

SPEAKER_00 0:16

You're listening to Unchecked, the podcast about the architecture of disinformation with Dan Brown and Rachel Price.

Rachel 0:30

Dan, before you say a word, do you watch the Gilded Age on HBO?

Dan 0:36

I watched season one and we've not picked up season two. Is it good?

Rachel 0:40

It's great.

Dan 0:41

I heard it's fantastic.

Rachel 0:43

Okay, so you know then that the Gilded Age, the actual era, is described as a period of flashy materialism and overt political corruption in the United States.

Dan 0:56

The name of the age gives it away. Yes.

Rachel 0:58

So roughly from like the 1870s to 1910s. I'm jumping right into my story because I'm excited. Yeah, I've been watching the Gilded Age, and my story takes place in the Gilded Age. Okay, so next question. Dan, what do you know about the Spanish-American War?

Dan 1:12

I know which countries fought in it.

Rachel 1:15

Good. I think that's really all we need.

Dan 1:17

And around the time it took place, I'm not a historian by any stretch.

Rachel 1:21

I literally have it in my notes here. Dan, I'm not going to explain this to you because I'm not even remotely a historian. Okay, so my artful context setting is now done. My story is about newspaper rivalries, sensationalism, and the Spanish-American War. So it's January of 1898. Pulitzer and Hearst are rival newspaper magnates at this time. And in an attempt to increase circulation of both papers, they are uh deep in this rivalry and competing via sensationalism in their headlines and in their writing. At the same time, tension had been brewing between the US and Spain for quite some time. The battleship USS Maine is sent to Havana during the Cuban uprising against Spain to just like keep an eye on things and protect American interests there. And in February of 1898, the ship sunk as the result of an explosion with unknown origins. It was really awful and really sad, and lots of people died. Also sad is that I bet you can guess what happens next. Misinformation, loves a vacuum. In the absence of information about what actually caused the explosion, who wrote some really amazing headlines? Hearst and Pulitzer's papers. Both just immediately went to blaming Spain for the explosion. These are headlines like quote, torpedo hole discovered by government divers in the main, startling evidence of Spanish treachery revealed. And another one that just says in all caps, Spain guilty. And so these newspapers then called President McKinley and the US military to take action to force a military response. Pretty soon we entered the Spanish-American War. And another newspaper, the San Francisco Call, satirically wrote, quote, the prompt and able fashion in which Hearst and Pulitzer took charge of the government at a critical period. So newspapers around the country were openly calling attention to and even satirizing the power that Hearst and Pulitzer had, and their newspapers had on the political actors of the time, including the president of the United States of America. Misinformation led us to this war. Now, this isn't the only story we've done with that takeaway. Dan, you did a story earlier about this same thing, and I'm sure there are many other stories like it. The reason I brought this story in particular is because I was really struck by the awareness of the media industry and thus many civilians who read these newspapers, but the awareness at the influence of these magnates of I was gonna say institutions, but I'm not using that word, of these businesses. The fact that this influence was even satirized just really struck me. Maybe it was a reflection of the time that these papers were so brazenly influencing major global events, but it kind of seems more likely that this has always been true. And yearning for some fictional time of days gone by when journalists or newspapers were stalwart and objective and neutral entities is actually very silly because maybe that has never, ever, ever been true.

Dan 4:54

What an interesting story. From actually this week that we're recording this, middle August. But earlier this week, the president held a press conference in which he indicated that he was activating the National Guard and federalizing the Metropolitan Police Department here in DC, which was honestly downright scary. It's been difficult for me to kind of consume and watch the press uh conference, but I've got some choice quotes from there. Towards the beginning, he said, I'm announcing a historic action to rescue our nation's capital from crime, bloodshed, bedlam, and squalor and worse.

Rachel 5:42

Bedlam. Sorry, continue.

Dan 5:47

He really likes that phrase, liberation day. And we're gonna take our capital back. We're taking it back. He said the murder rate in Washington today is higher than that of Bogota, Columbia, Mexico City, and some of the places that you hear about as being the worst places on earth. Much higher. This is much higher. I have traveled into downtown DC many times.

Rachel 6:10

And survived.

Dan 6:12

And survived. Lived to tell the tale, despite the bedlam and the murder. And the last thing I pull out is he said, our capital city has been overtaken by violent gangs and bloodthirsty criminals. I guess of vampires. I don't know. Roving mobs of wild youth, drugged-out maniacs.

Rachel 6:32

Wait, are the drugged-out maniacs separate from the wild youth?

Dan 6:36

It appears to be a list of Oxford commas. So yes, separate. And homeless people. Anyway, it is incredibly frustrating to read this characterization of DC. I mean, I grew up in New York City in the 70s and 80s. I know, you know, bloodthirsty. I know bloodthirsty. I've experienced it, I've seen it, and I haven't even seen the worst of it. But I understand that cities, when you pull a lot of people together, can be, let's call it complicated uh places. But Washington itself has its problems, we know, but it's beautiful. It's it's a beautiful town. I like to think that uh the local governments here are working to address these things. So I didn't want to sit here and kind of dispute these. He actually spews a lot of so-called facts, and actually PolitiFact did a really good piece highlighting the inaccuracies of the statistics that he quoted. Reality is crime is way down in the DC area, and the ways he's describing it and the statistics he quotes are are woefully inaccurate. But there's sort of three disinformation things that I wanted to pull out of this that I noticed. First of all, we've talked so much about framing, and the framing here is just such a classic example. Yeah. The frame that I took away from this, maybe you took a different one away, is it's either bedlam and violence and bloodthirsty criminals, or increased police presence, federalizing the police force, adding the National Guard. Like those are the two choices.

Rachel 8:12

The framing I noticed was taking the city back implies that someone has the city under their control, right? Or like the city must be retrieved from an enemy's force.

Dan 8:26

Whoever's taking it, yes. The bad guys who took it, yeah. Totally. Obviously, cherry-picking data, right? It turns out, and again, PolitiFact does a really good job breaking this stuff down. He's picked cities where, yes, the murder rate is lower, but there are cities in the world where the murder rate is much, much higher than the ones uh that he picked and of DC. There's cities in this country where the murder rate is higher than that of DC. So it sucks. I wish there were zero murder, but to kind of frame it in a way like DC is some kind of dystopian horror is terrible. And then the last thing that I just thought of when you were talking about the Gilded Age is that the press reports the statements that he makes as the news, not the inaccuracies as the news. That we have political fact, you know, to kind of do some of that fact-checking. My understanding was that during the press conference, CNN, I think, was broadcasting the actual statistics online. So we did get some of that, but for the most part, most people are going to just read the fact that he said these things as the news and not the fact that he was wrong and what the actual statistics are. Back in April, the Justice Department put out a statement, said that crime in DC is at an all-time low, and it's was because President Trump was in his first hundred days. So they have both taken credit for the all-time low crime and disputed the all-time low crime.

Rachel 10:01

Yep, they yep, that's true.

Dan 10:03

We're gonna move on to something much more interesting now. We're gonna talk to Claire Wardle of Cornell University. Anything you're looking forward to in this conversation?

Rachel 10:14

Oh, Claire is such an engaging person to talk to and a solid comedian to boot. So this was a really, really fun conversation. Today we're speaking with Claire Wardle. Claire is an associate professor in the Department of Communication at Cornell University who studies information ecosystems with a focus on effective communication in an age of AI and conspiracy. I think we might get into that today. Claire, welcome. I think we will too. We're just so thrilled to have you. Claire gave the keynote at the information architecture conference in 2025 in Philadelphia. Dan and I were both there and both had so many questions and we're so eager to talk to you. So thank you so much for joining us. No, my pleasure. Thanks for having me. When Dan and I started this podcast project, we were pretty eager to define misinformation and disinformation. I think I even found this wonderful resource you wrote up defining all the different kinds of mis and disinformation. And it was a great primer for learning about the space. I've recommended it to a lot of people. But the further along we get, the more we've realized how challenging it is to actually define all of these things. And you start to wonder if it's so hard to define these things, like should we be defining them?

SPEAKER_01 11:43

Ugh, so I think definitions are always important. I'm an academic, of course, I'm gonna say that. But I think there was a rush to define these topics, sort of 2016, 2017, because everybody was suddenly fascinated by it. And I think we thought, well, if we can make what's so complex easier to understand, then we can be like, oh good, we've done something about it. But to your point, I think the more that you actually play with this content, and I always jokingly refer to spending my days in the armpit of the internet, when you spend all day looking at the reality of what's there, those simple definitions don't make sense as much, right? And so I think some of the things I've reflected on as somebody who really lent into the importance of definitions is I think that I didn't help the field by trying to make it simpler than it actually is. Because people were like, oh, here's an example that fits that nice typology and this idea of, well, misinformation is people who don't mean any harm, and disinformation is people who are deliberately doing it for financial gain. And yeah, that had value to help people think about some of those differences, but I've increasingly spent time in spaces where people are almost in a completely different reality. And for example, anti-vaxxers really deeply believe that by not speaking up about vaccines, that that would be causing harm, that vaccines are causing harm. And so they need to push this information. They're not like my mum failing to verify something she saw on Facebook. They're in a completely different reality, drawing completely different conclusions when reading the same scientific articles that I'm reading. And so those simple definitions from 2017 just don't get at that. And so I think for those of us who are working in these spaces, designing interventions, trying to improve our information ecosystems, we have to lean into the complexity as opposed to pretending that it's more simple because it's not.

Rachel 13:26

So, what does it look like for us to lean into that complexity? Like, what does that give us?

SPEAKER_01 13:30

So I think we have to recognize that the focus on content, like, is this true or not? Should we label this as true or not? Doesn't get to the psychological drivers that influence why people share something or consume something or put it on their account as a badge of honor, right? There's so many ways that information is connected to people's emotional drivers and their identity. Like when I make decisions about what I share, I'm thinking, who's my audience? What do I want my audience to think about me? What community do I want to be seen to be part of? Is all that stuff going on? And I'm not thinking, well, maybe I am sometimes because I'm an academic, but this idea of like, oh, the label tells me it's false, therefore it must be false because those people know what they're talking about. We know that that's not the way that people navigate information ecosystems these days. But it's hard, I think, for those of us who have been trained in facts and data. To us, that is the most important thing. And if only we just told people that it was true or false, if only we had more facts, more diagrams, more PDFs, everything would be okay. And the truth is that's not actually how people have relationships to information. I think there's a lot of discussions like, oh, it's a problem of the right. This is a human problem. Every one of us has our own biases, every one of us has an emotional relationship to information. And I think those of us who help design these spaces think about the quality of information. We have to let ourselves go a little bit. We have to kind of get rid of some of our training to say, hang on, what would it mean like to lean into emotion, to lean into stories, to lean into words like we as opposed to, or I, as opposed to a distant passive voice, third person voice, right? But that it feels uncomfortable to many of us who are working in these spaces.

Rachel 15:16

One of the topics that's come up for us, I think a couple times throughout this podcast series is one of the proposed solutions to misinformation is to just release better facts and just put more, quote, good information out there. And then that will, you know, somehow cover up or replace the quote bad information. And I think when you sit with that idea for more than 30 seconds, you realize it's problematic. Like if that worked, we would probably have solved it by now. But what I'm hearing you saying is agreeing maybe with that, like, yes, it's it's not gonna work to just release more facts because we really have to think about the psychological drivers of why misinformation thrives, why people share it, why people are even engaging with it at all, if we have any hope of addressing the overwhelm of disinformation we're experiencing right now.

SPEAKER_01 16:08

Do I have that right? Yeah, you do. And I think again, coming back to our own experiences, all of us at the middle of the night have been worried about some health issue, either our own or somebody we love. And we go to Google, right? And we get to like an NIH website or a CD, and you we read the facts, and you're like, okay, I'm just gonna check Reddit to see if I can find somebody else who's experienced this. And then you go to Reddit, and then you find yourself on TikTok with somebody talking at you whilst crying, saying this happened to them and they did this as a way to do something about it. And you go back to sleep. What's the thing that resonates? It's probably the person on TikTok, maybe the person on Reddit. The CDC, it's not resonating, right? Yeah, and that's what I say to those of us who are working in this space. If you're scared, if you're frightened about your health or somebody you love, facts don't matter at that point. You want some feeling of control. And that's why we're seeing all of these really problematic health interventions. I don't, you know, people have watched that Netflix documentary, Apple Cider Vinegar, right? It does a really good job of explaining when you have had a very serious diagnosis, it doesn't matter what the doctor says to you, you're being told something that's very difficult to hear. And you want to bring some control back in your life and being told that there's a simple formula that's going to help and going to work more quickly, of course you're gonna go in that direction. You know, and I think we need to sit with that longer than we currently are. But what that means for our design, I mean, I'm not saying that public health organizations should get into memes and should be on TikTok doing wacky videos, but it does mean how can we be more honest? How can we be more humble? How can we be more empathetic? How can we think about information in ways that aren't just facts, facts, facts, dry gram, dry PDF facts?

Dan 17:44

Do you see anyone doing that well these days?

SPEAKER_01 17:47

I see some really, really impressive um healthcare workers in particular on TikTok. And I did some research with a student about misinformation related to RSV. And what we realized that it wasn't as much doctors doing their own videos, it was doctors who were getting into the comments of videos that were sharing misinformation. Just been like, hey, just to let you know, and like being part of the conversation, not being the like, I'm broadcasting to you. Like, really, I want to hear what you have to say. I understand that this is scary. You know, as somebody who has treated thousands of people in your situation, I'm just gonna share what I've observed, like leaning into observations as opposed to like this is what the latest Lancet article says about RSV. Right. So I think what I see is I don't have many organizations that I think are doing this well. I see individuals who just have a natural talent for listening and empathy and engagement. And when you see it done right, you're like, oh yeah, that's how it should be done.

Dan 18:40

There's this guy I follow on Instagram. He's a young guy, he's based in North Carolina. I think his handle's like Dr. Nock or something. He's like a biomedical researcher, but he's he puts out these videos that are very informative about the stuff. But his videos are so well polished, people accuse him of being an AI because he looks I don't know what happened. Like he's sort of like this normal guy who's really into science and really tries to make things accessible, but it's still backfired on him. Like he's still just one guy looking into a camera, and yet because everything looks so nice and because he's so, I don't know, clean cut, I don't know, people are accusing him. He's got really nice skin, too.

unknown 19:19

I think that's a big problem.

SPEAKER_01 19:21

But you've got at something really critical, actually, which is in this moment, people are seeking to use that terrible phrase authenticity. And unfortunately, because if you don't look a bit scraggy and you don't look a bit unpolished, then people accuse you now of being AI. It was a problem we would not have considered five years ago. But now people really want you in the middle of the night. You know, there was no AI involved. I mean, if you don't look tired, then you must not be of this planet.

Rachel 19:48

I would like a bumper sticker that says that. Great idea for merch. That'll be our first uh podcast merch. Yeah. I heard you talk about joining the conversation and this word engagement. And I think we're kind of dancing around this idea of the medium of the content and also the delivery of the content. So you've raised the idea that a lot of disinformation tactics take advantage of the internet as a medium, right? Thinking of these late-night TikToks and that sort of thing. When it comes to misinformation, do you think the medium kind of is the message?

SPEAKER_01 20:24

Yeah, I mean, I'm a communication scholar, so give me more material. But I think that is absolutely right. I mean, I talk to organizations and say there was this thing in the early 2000s called Web 2.0, which was the idea that it was no longer flat on a website that people could talk to you. You did not listen to that shift and you continue to just broadcast on your website, or if you do tweet, you say, click on our link, which links to a PDF, or if you're a news, you like tune in at 11, listen to our webinar. But the idea of actually having a two-way conversation, which is fundamental to the internet, I think is forgotten by too many people. And I think that sense of medium is the message. If somebody's giving you advice that looks like you and they're listening to you and you're having a conversation with you, like that's built into the message, you are gonna have a different relationship to that than you are the top-down, polished PDF or the broadcaster, whatever. So that element of like the design of so many UGC platforms, I think has got to this sense of like it looks authentic because there are people who look just like me. And that has happened at the same time as the decline in trust in institutions, right? So we can't separate the two things. Like the global financial crisis, nobody ever said sorry. Like we've had a whole host of issues subsequent to that, where people have lost trust in a whole host of institutions from banking to politics to news, et cetera, et cetera. So that has happened at the same time as we've had the rise in these spaces where people who look just like you tell you that you're right and are listening to you, which is, I'm sure, for another question about like Chat GPT, right? The design of these LLMs is to be conversational. And to tell you that you're right. Oh, lo and behold, people are becoming completely seduced and believing all sorts of things. Right. So we do have to think about the design of these structures and how that builds trust in an era of decline in trust in institutions as we've known them for hundreds of years.

Rachel 22:16

You took me there, so I'm going to follow you there. What about LLMs makes them this perfect media for disinformation?

SPEAKER_01 22:24

I mean, it doesn't help that the fact that it's sort of baked in that sometimes it will just tell you something that's not true. But the thing that's kind of fascinating about hallucinations is it tells you it with real confidence, right? Yeah. The other day I asked it, I was like doing all this research, and I said, I can't find any qualitative literature on X. What am I missing? And they gave me three responses. The top child was like, oh, I can't believe I didn't think of it. The third one I was like, oh, that sounds great. Of course, go looking for it, it's not there, right? But the confidence by which it told me about this book was like, oh, okay. But it works because it sounds really authoritative. It looks professional. I mean, even the font, like it just compared to a website with pop-up ads that doesn't load and has a paywall, there it is, looking sleek, looking great on your phone, loading quickly, being polite and talking to you with empathy and telling you, you got it, Claire. What a great idea, Claire. Like it just, it really is seductive because it draws in on all of those cognitive elements. So to go back to the beginning of the conversation, what have we not been designing for? We have not been designing for empathy, conversation, this sense of you're part of something, you're part of a community. And LLMs came along and be like, oh, we got this. We're gonna take all those things that you haven't built and built them into this machine, and we're gonna be there for people on their devices, talking to them and making them feel heard. Oh, surprise, surprise. Really seductive and effective.

Rachel 23:48

There's an exchange there. I'm thinking about how you know we make sense of spaces and we make sense of information, and it's never from a blast announcement, and I read it and now I fully understand. There's this exchange of asking questions. Yeah, it's a process. There's asking questions, there's getting clarification, there's thinking through things. And LLMs are the space where you can do that. Yeah. And receive information, whether or not that information is accurate. But that space is a space where you can do that.

SPEAKER_01 24:21

Yeah. We can joke about it, but the anthropomorphism of saying, Would you like me to see more of this? And I type, yes, please. Yeah. Right. And you think about traditional institutional websites. I know it sounds silly, but like we haven't been very polite to people. Like we're like, here are the facts, believe us, trust us, as opposed to the way that it brings you in, and you end up being polite to this machine, even though I know it's a machine. That to me is incredibly worrying, but also impressive.

Dan 24:48

Yeah. As you're talking, I'm sort of picturing this timeline of toxicity, um, which is the only way I can think of it. But you were saying that, you know, when we looked at social media, we see people who look like us. I wonder if there was something, a transformation that happened there that's even more insidious, which is where content creators weren't people that looked like us, but were looked like people we aspired to be. And that as they provided advice, true or not, accurate or not, because we aspired to be more like this person in this lovely place, looking as good as they do, we bought into what they were saying. And now the transformation that we're seeing is, and I do think this is an illusion, but it feels like the LLMs are not just being polite, not just being affirmative, but seem like they're actually listening to us. And there's something about being heard. There's maybe some main character syndrome baked into this too, but right, it seems like someone is actually listening to me on the other side of this screen is very, very tempting. I think it draws people in to the point where they'll believe anything that comes up on the screen. I didn't have a question. I was just making an observation.

SPEAKER_01 26:04

Well, but I do think it gets to, you know, as somebody who spends a lot of time in conspiracy spaces or disinformation spaces, what you realize is A, a lot of this intersects with our loneliness epidemic, because you would see people like QAnon believers who would say, you know, I work a night shift at one o'clock in the morning, nobody else is awake, but I'd log on to QAnon and there would be somebody there being like, Hey, how are you? Right. And I think this idea of feeling heard, again, if you spend time in these disinformation spaces, there's not hierarchy in the traditional sense. It doesn't matter who you are, you have the right to do your own research, to use their phrase, and you can share that research, and that research will be listened to and used by others. So you feel like you have a role to play. We're not being heard by the New York Times. Like I don't feel part of the New York Times. Like the other day, for the first time, the daily said, we asked for some listener questions, and we're gonna answer them as if they were like this amazing thing. It was like 2025. I was like, oh, well done, the daily, right? But like we don't feel heard by institutions and they're not really even bothering to listen. The sadness is the kind of the absence for many people of human connection means that even just a connection with a machine is enough. And you see people now very quickly having relationships with these machines because it mimics what they used to have.

Dan 27:19

What should we be on the lookout for, do you think? Like if you were just to speculate, I won't hold you to anything, but I'm just really curious. Like, as we are building our timeline of toxicity, like what's the next set of behaviors that these systems, these algorithms, will create to draw us in even more, do you think?

SPEAKER_01 27:37

Yeah, I mean, I think the amount of money that's being invested in these technologies means as much as there's concern about them, they are just the money's just going to flow in their direction. And so what we see is I think there will be more and more personalized agents. And so all of us will have a personalized agent that we might pretend is the thing that we ask to book our flights, but in the middle of the night, we're like, hey, I've got this mole on my leg and I just wondered if you could have a look at it for me. But also, hey, I'm really busy today. Can you just give a snapshot of three things that I need to know, right, as I'm driving to work? And what we won't realize is when we are concerned about echo chambers and filter bubbles, right? Like this is that on steroids. The idea that it's gonna feed us more information, that it knows better than any previous algorithm what I like. Here you are, Claire. Because it knows absolutely what I'm worried about, what's on my mind, and it's gonna feed that because it wants me to come back the next day, right? Because it's all the same stuff. So I think Dan, the timeline of toxicity is it's gonna take us a while, I think, to recognize how personalized those agents are. Because we're not even with our communities online, being like, oh, my best friend or my neighbor is also here. It's like, oh, I assume everybody's hearing this every morning, but that's not the case. We don't even have any visibility into what everybody's searching for. Like on Google, there's kind of Google trends, and you can be like, what are people worried about? Oh, everybody's Googling Sydney Sweeney this week, right? But we don't have that with LLMs. I don't know, is everybody prompting LLMs about moles this week or wildfire smoke or RSV or like who knows, right? So the other thing is we have absolutely no visibility into what it is that people are going to consume. So not only can we not intercept, we can't even know.

Rachel 29:14

And the deep skeptic in me, you said, who knows? I'm like, some people get to know it's not me. Yeah. Dan, are you okay? Because the face you made about a minute ago scared me.

Dan 29:26

I have not been scared until Claire, you started talking about those personal lines. Like I was definitely having a physical reaction to thinking about how insidious it could be.

Rachel 29:40

We were talking with Jave Dylan Thomas about confirmation bias and operationalized confirmation bias. And this is that on steroids.

Dan 29:49

Yeah, 100%.

Rachel 29:50

Where it's not only that I'm only going to pay attention to the things that confirm my worldview, it's also I'm only going to be served the things that confirm my worldview.

SPEAKER_01 29:59

But we won't think that, right? Like in Google, you can see all the blue links and you know that you're heading to certain sources. You don't get that with an LLM. Yeah.

Dan 30:08

I feel like we have people in our lives. For some of us, it's a spouse or a life partner. For some others, it's just a you know good friend or whatever. But you think of something in earlier today and you're like, oh, I'll go tell that person about what I was thinking about. And they're they're interested because you're you and you have this relationship with them. And in this world that is so crowded with information, I sometimes find myself tempted to not share those things with my wife because it's like, well, she's had a bad day, and it's been she's been really busy, she got a lot on her mind, there's a lot going on in the world. I'll just keep that to myself, no big deal. But like the more we do that, the more we will, I think, seek an outlet for those things that will not just listen, but like be a willing, enthusiastic listener. I can say stuff to my lovely wife, and she will pretend to be interested. But I know it's like some nerdy, weird UX thing that I'm thinking about, and she's just, we're on our dog walk. It's fine. She's listening to me, I'll listen to her. But at some point it's gonna be like, no, I don't want a little bit more. And so I'll turn to this personalized agent who'll say, Hey, I know you were working on that design problem. How'd you do with that? Yeah. And it's like, thanks. I'd rather talk to her, I'd rather talk to my lovely wife, right? That's what got me scared when Claire was talking. I was sort of thinking about these insidious opportunities to kind of create even more separations between people. We see it happening kind of writ large, but it's going to become, I think, maybe more intimate too.

Rachel 31:34

Where I worry about that is I've got a three and a half year old at home, and I'm really worried about the act of interrogating a new idea and learning more about the world and trying to learn about a thing is really exciting, right? For an adult and for a kid. And I think, especially for a kid, like learning how to find the information, like learning how to answer the questions that you have about the universe is thrilling. And there's absolutely a role technology should play in that because technology can provide access to so much information. The thing I worry about is I think there's also a very human kind of bonding component to trying to investigate something new, getting help, figuring out like where to look or how to interpret the information, and then like sharing that excitement with somebody, especially like I'm assuming there's a positive developmental impact to doing it not in isolation. And then when we look at how easy it is for a person to do that kind of interrogation with an LLM in complete isolation, that freaks me out.

SPEAKER_01 32:38

And I think because I obviously teach a lot of students, particularly undergrads, and my concern is their use of LLMs because it's frictionless and easy. Yes. Like Sherry Turkle has written a lot about it's not as if you can't get something from having a relationship with these characters, but like having a relationship with a human is hard. Like sometimes humans are, excuse my language, dicks. Like sometimes they don't respond to your call or they're grumpy, they've had a bad day, whatever. But like LLMs never get tired. LLMs are never going to be rude to you. And she talks a lot about like grief bots. And there was a thing this week with Jim Acosta when he was talking to an AI version of one of the kids who lost their life at Parkland. He's about to turn 25, and the parents have created an AI bot of the son. And Jim Acosta interviewed him about why do you think part of it? And like Sherry Turkle talks about grief in the sense of like, of course, we might want the bot, but like grief is hard and you've got to go through that pain. And I think about that all the time. Like you're talking about your child, like learning. Like sometimes learning is deeply frustrating, and like you have to go to the library and find a number of different books. It's not easy. LLMs are like, here's my question. Oh, here's a paragraph that I've given to you in a split second. Oh, thanks. Like, yes, there's benefits to all of that knowledge, but like if we don't have the friction and the pain, yeah, and that's what I see with my students, right? They just want a completely frictionless life with their relationships, with their schoolwork. You know, and so in that space, they're turning to what's easy. And then when I any of us ask them to do something hard, they haven't got the resilience because they've been so used to this frictionless environment. So that's the other thing I worry about here. There's filter bubbles, but there's also what does it mean to lose the stuff that's difficult?

Dan 34:13

Um, I value the friction.

SPEAKER_01 34:15

Yeah, because you've been taught to value the friction.

Dan 34:16

Because I've been taught to. How do we help people understand the friction feels in that moment like it sucks, but in the long run will make you a better person.

SPEAKER_01 34:26

Well, they're all into weightlifting, young people, right? So it should be like, at the time, it's like it hurts, but you're creating these little tears in your muscle that in the end build a lovely strong like we have to do something to say no pain, no gain. But when it's so easy to ask the LLM, I mean, we'd have to take the devices away from them.

Rachel 34:46

Well, I think the other thing too is this isn't a bunch of individuals making a choice to take the easier out because they're lazy. This is a bunch of individuals who are existing in a system that has taught them deeply that speed, efficiency, and productivity, neoliberalism, delivery are output. Thank you. Yeah. And so, of course, if you've been taught to optimize for output and you see a tool that speeds up your output, hell yeah. Yeah, yeah, yeah.

SPEAKER_01 35:17

And also, we have to go back to Silicon Valley designing things that make it really, really easy to use, like the affordances of these platforms, which are frictionless and seamless and load quickly. And you know, at the L end of an LLM, it says, Could I possibly do that for you in this way? And you're like, Oh, thanks, hadn't even thought about that. Like they're keeping you on there all the time. So we can't also we have to recognize that we as humans are being seduced by the design of these technologies. And to go back to the very beginning of this conversation, you know, as people who are thinking about information design, we're out here being like, should we have more labels on it? Like we're trying to build in the friction to the system. They're over there creating beautiful ski slopes of like, there's no friction here. Yeah. Like just keep spending time on our platforms, right? And so, how do we think about friction in our designs, but not doing it in a way that emphasizes the bad stuff but creates frictionless environments for the good stuff and enjoyable environments.

Rachel 36:14

Yeah, when when you go to the IA conference, you're not gonna hear too many people talking about designing for delight. So we've been talking a lot about like current state and future, but Dan and I also like to kind of look into the past, right? Because we know misinformation is not new. And so thinking about the internet, the internet seems to be on this arc to make it more and more conducive for misinformation. I'm curious, what are some of the seismic shifts we've seen over time that have helped misinformation thrive? Because I'm thinking LLMs is not the first one.

SPEAKER_01 36:47

No. I mean, yes, misinformation has been around since the beginning of history, right? Like there's this great study where people looked at how gorillas and apes connect with each other and they pull off fleas from each other's fur, but we don't have fur, so we gossip. So I love to remind everybody about like we've really so-and-so sleeping with so-and-so? Oh, I haven't verified it. I don't care. I'm just gonna share it because it's so great. But for me, who's been studying this stuff for 20 years, like really Photoshop was the first kind of device where it was like, oh, you can have an eight-year-old with a computer who can suddenly create something very different. And I started this work back in the day at the BBC, and they would often get hoaxed or they would get sent something that was photoshopped. It was like, oh, how do we look at the exif data to analyze whether or not this photo is real or not? So we've kind of had those sorts of things, but then the ability for anybody to create a website. The first use of the term fake news was kind of 2014, when people were creating what looked like newspapers, but they weren't newspapers and they were online sites. So that was a big thing. And then in many ways, just things like the visual turn of things like Instagram and now TikTok and YouTube, like the ability to edit. So again, the technology wasn't new, but anybody could use it. So the democratization of technology meant that anybody could use it. And even for those of us who study it, right? It's much easier to study text. So a lot of fact checks and others are on text. It's really hard to fact-check YouTube and TikTok or podcasts. We don't have machine learning abilities. Really, we do a little bit with podcast transcripts. So there's also been this kind of arms race around technological shifts allowing this information to thrive, and for those of us who are trying to fight it, making it harder. So again, when I was back at the BBC in like the early 2000s, the BBC would get hoaxed like three times a year, and now they get hoax like three times a second, right? Like the astronomical shift in how easy it is to create this and how easy it is to disseminate. So often when I do talks, there'll always be somebody at the back being like, oh, are you telling me that misinformation is really new? It's not new. It's like, no, it's not new. The psychological drive, none of that is new, but the ability for an eight-year-old with an iPad to create something so compelling and the ability for algorithms to get into people's faces in seconds, that is what's new, and that has driven this huge, huge shift, which there was never an internet that was perfect. Obviously, we know that. But also, we're never gonna get back to an internet that is better. For me, we've been disproportionately focused on kind of the 20% of the bad stuff, but actually, how can we focus on the 80%, which has been pretty mid-level? Right? We try, Claire. But I but that to me is that that's the bits that we can control. Like, you know, there's always going to be financial factors that lead disinformation agents to create shitty content, but there's very little we can do about that. What we can do is make the rest of the information ecosystem more engaging and joyful as well as factual, and that's what we should spend our time on.

Rachel 39:42

You are in a lot of different spaces talking about misinformation, having a lot of conversations. Uh, I'm curious at a meta-level, what are some common misconceptions that you hear about misinformation? So maybe misinformation about misinformation.

SPEAKER_01 39:58

It still feels pretty stuck where the conversation was in 2016. And still in conversation after conversation, there's this idea of like we just need more fact-checkers. And I love fact checkers and many of them are my friends, but like that's not gonna get us out of this hole. And I think it goes to the beginning of this conversation. The misinformation about misinformation is that we can fact-check our way out of this, that we just need higher quality facts and everything will be okay. And in communication theory, we have this theory called the third person effect, which is that individually, none of us think that we're affected by advertising, but everybody else is affected by advertising. And it's the same with misinformation. It's like, oh my goodness, I've never been affected, but everybody else has. Day three of the pandemic, I was living alone in a studio apartment in New York City and forwarded a text message that said the city was going to shut down, there'd be no toilet paper. And five minutes later, I was like, I'm the woman who teaches this stuff. Yeah. But I was scared, right? Like, and so it's like every one of us, doesn't matter how much media literacy training you've had, right? And I just wish there was more of that because we're stuck in conversations that still feel 10 years out of date.

Rachel 40:57

So this has all been like super upbeat, really, really joyful stuff. What keeps you optimistic? Like, let's let's try to end on a high note here.

SPEAKER_01 41:07

So fundamentally, the fact that technology shifts as quickly as it does makes this stuff endlessly fascinating, right? Just as I'm like, you know what, I'm kind of done. And then LLMs come along and I'm like, what? So I think just from a kind of academic curiosity, that's always fascinating. But I do think the joy for me is what we do have control over. Like, I do think there are opportunities to do really good work, listening to communities, engaging with communities, meeting communities where they're at, answering information needs. Like, I think there's all these opportunities that aren't actually that difficult to do. We just need to do them. And so the thing that sparks joy for me is doing a community engage project or actually listening to people like, you know what, it wouldn't take a lot for me to improve trust with X. I just wish this or whatever. And, you know, some of the early projects in the early 2000s that involved crowdsourcing and really finding ways to tap into lived experience. Like that to me is what's amazing about this space.

Rachel 42:05

Claire, thank you so much for joining us. Uh, this was a really fun conversation and we learned a lot. Thanks so much for having me. I loved it.

Dan 42:18

I was really excited that we got to talk to Claire because she gave the keynote at the IA conference, and the moment she started talking, I was enthralled and enthused by her manner and her approach to this topic.

Rachel 42:38

It's so telling. One of the things she talks about is how we need to deliver information in a more compelling way. And she not only speaks to that, but she actually acts that out in how she is delivering information. Yes. Uh, so it's really fun to see that.

Dan 42:57

Something else she said, and that's what's subtle, but I really appreciated it. She said a few years ago when people became a lot more aware of misinformation, she put out a guide to misinformation and she sort of wondered aloud if maybe she had done more harm than good. And I think that is such an important part of being sensitive to disinformation is acknowledging and recognizing harm that you might do. I don't claim to know whether she did any harm or not, but just the fact that she brought it up and thought about it. There are folks in the public who would not do that. And I think you need to be more suspicious of people putting out information that refuse to correct themselves when they've made a mistake.

Rachel 43:40

Yeah. It was funny that she mentioned that source and wondering if it did more harm than good, because I realized later that source was one of my entry points into this space. So when you asked me to do this podcast with you, and I said yes, and then I went, oh no, I don't know enough. And I went and did this whole like Literature review basically, and the primer she wrote defining the various kinds of misinformation was hugely helpful for me. But one of the things we talked about a lot is how much time do you really spend defining this stuff? Right. And that one I felt both seen and attacked as an information architect. I love defining things. But there is a point of diminishing return on doing that.

Dan 44:24

100%. What lens did you arrive at?

Rachel 44:28

Today's lens that I want to bring is friction.

Dan 44:31

It's a classic.

Rachel 44:32

It's a classic. I don't think we've actually we've discussed this lens before in the context of our interview with uh Sid. I don't think we actually brought it into the episode.

SPEAKER_00 44:43

Nope.

Rachel 44:44

So the lens of friction is wondering how does the system insert the appropriate amount of friction in the process of contributing, transmitting, interrogating, understanding, or distributing information? And wondering how does the system introduce friction without putting people off or with putting people off? You know, I don't want this lens to imply there's a correct way to apply friction because I think it's so contextual, but it does ask you to explicitly think about the role of friction in the production and the digestion and the spread of information.

Dan 45:23

One of the things that I've observed about these conversations that you and I get to have is that this is put totally putting you on the spot. You are really good at thinking of examples of things. So I wonder, do you have an example of friction in the wild? Good friction.

Rachel 45:39

Good friction. I don't know that this has to do with misinformation, but the top example that comes to mind for me is when you send an email in Gmail and you say that you're attaching something, and then you hit send, and the system is like, you said you were attaching something, but nothing is attached.

Dan 45:57

Right.

Rachel 45:58

Do you want to send this? I think that is a good example of friction.

Dan 46:03

Yes.

Rachel 46:04

When I think about friction, I'm actually thinking about a way to rough up the edges of the information delivery so that you are not too comfortable, you're not too complacent, and you do remember to question things.

Dan 46:20

Yes.

Rachel 46:21

There's this kind of de facto assumption that as designers, we should be designing these delightful, easy, dare I say, frictionless experiences. And I think that frictionless is not the right direction. Right. I think there are moments where you do want to intentionally introduce friction. This is not a new idea. I feel like we talk about this a lot, and this lens is really trying to elevate that back into a conversation. This feels related to me to as few clicks as possible. Oh, yeah. You know, when I'm designing uh workflows, user flows, whatever, there's this constant assumption that we should be shortening the path and making it as few clicks as possible. And I think in a lot of situations, that is not correct. Right. There are moments where you want to introduce friction. I think where we see friction required is in regulated areas. Banking is top of mind for me, where you actually don't want a completely frictionless experience because the outcomes could be disastrous. When you are paying someone on, I believe it's Venmo, especially the first time you pay someone, you can look up their contact and you can just be like, great, here's 50 bucks. And there's this moment where then you have to enter the last four digits of their phone number. That is friction to keep you from accidentally sending money to someone you didn't intend to send money to. So that is another great example of friction.

Dan 47:53

Those are great examples. And they're great examples because, like you said, this idea of friction has been around for a while in UX design. But the intent of that friction is to prevent someone from making a mistake. And now we're looking at the lens of friction through the lens of disinformation. What role can friction play in creating systems that, let's just say, sort of help people understand what they're looking at, is true or not true.

Rachel 48:23

Yeah. Dan, what was your lens?

Dan 48:26

It's my turn. I'm I don't have a good name for this lens, but she told us a fable. Let's call it a fable of the person who wakes up in the middle of the night dealing with something. It's maybe something medical that they're worried about. And yeah, they could try and dig through WebMD. Yeah, they could kind of go through the phone tree of their doctor's night call hotline or whatever, but they end up on TikTok. And there's someone who's staring right in the camera, who is providing a empathetic and real-world reaction experience to something similar. And there are reasons why that connects with us and why that sort of bubbles to the top for us of how to deal with these complicated situations. I feel like information architects have long thought about extremes, extreme cases, what we used to call edge cases, right? This sort of idea that we can't just think about, again, what we used to call a happy path, but instead think about some of these scenarios that might creep up. IAs, for some reason, are particularly good at this skill. But thinking about this, I've been calling this lens the middle of the night lens. And the way I pose the question is how would the system fare if it catered to someone using it in extreme scenario that prioritized efficiency or relatability or reassurance, right? If the role of the system is to provide a modicum of reassurance, if the role of the system or the content in the system is to provide some sort of like temporary fix, how can the system support that need? The middle of the night lens, for lack of a better term.

Rachel 50:14

In the middle of the night. That's all I can think about right now. No, I'm with you. I'm here. I love this lens. What I love about it, first of all, is that it's related to a bunch of other stuff that has come up around dealing with arousal states, extreme emotions, right? Or potentially exploiting them. Stuff we've talked about around vulnerability. This really strikes me as a like, how does your system attend to or not vulnerable states?

Dan 50:44

Yeah.

Rachel 50:45

Claire talked a lot about us thinking about the psychological context in which people are receiving and sharing information, and this really gets at that. And I like your name for it because middle of the night is a very distinct, it's real.

Dan 51:00

We've all been there.

Rachel 51:00

Yes. Can you think of examples of when this lens might come up?

Dan 51:05

So I'm thinking about when do I feel most vulnerable? That's a very practical framing of it. When your user is feeling vulnerable, they are heightened sense of arousal, all they need is information out of the system. For me, just to get very personal for a second, because I have a very gravelly voice, my voice doesn't do well with, you know, the automated phone trees where you have to speak in order to kind of move on to the next thing. I can't even say yes or no in a way that those systems recognize. Now, I may be overly sensitive. I know I'm overly sensitive about the quality of my voice, because I've talked to other people who don't deal with the same thing that I'm dealing with, and they also have problems with those phone trees. But because I'm so self-conscious about my voice, because I'm worried about how it sounds and how it's being received, I get stressed when I'm misunderstood by those robots, by the clankers. So my stress level goes up, I'm feeling more vulnerable. Usually I'm calling for something important, right? Because I'm dealing with a phone tree. And when you get stressed, your vocal cords tend to kind of tense up a bit, which makes the gravel in my voice even more pronounced. So when you were talking about vulnerable users, that is the first thing that popped into my head. How can a system like that that has important information in it, that you're gonna force me to go through one of these trees for, how can it make it easier for someone who is tense and stressed and whose voice quality changes? And that's the only way to navigate the system.

Rachel 52:42

That's a really great example of like a temporal interaction that kind of ratchets up with stress. One example I was thinking of when I think to middle of the night, I think baby and fever and first-time parent. Me.

unknown 52:56

Yes.

Rachel 52:57

A lot of us have been there. One of the things I've come to appreciate about one of the sources of medical information online, I honestly can't remember if it's like Cleveland Clinic or Mayo or one of those. There's this section in every article that is like when to contact your doctor.

Dan 53:13

Yeah.

Rachel 53:14

And so there's all this stuff. You can freak yourself out with all the symptom explanation, all the like why is this happening, all that stuff, or you can jump right to like how do you know when you need to throw up the flag and go get help? And that is such a kind of subtle structure of their content that is profoundly supportive of kind of cutting through the noise for a potentially vulnerable person and doing a little bit of triaging for like how worried should you be. Luckily, I have a lot of friends who are medical professionals and they do the same thing to me. You know, they're like, I will talk to you about this, but first is X, Y, and Z happening. Right. Like, let's cut to that first. I mean, it's triaging, right? You're like checking for like, let's get a sense of where things are at, and then we can go back and kind of look at all the other details.

Dan 54:05

I'm gonna give one more example because it literally just happened to me this morning. I noticed a fraudulent charge on my business account. And I went to my bank, and they have a link right there that says dispute charges. Actually, this illustrates both of the lenses that we just talked about because there was friction. It said, Are you sure you want to go through the bank to do this? Don't you want to go through the vendor to do this? Like the person who who did the charge. It said, usually you can resolve these disputes a lot faster if you go through the vendor. And I was like, that's an interesting thing, but I think this is just straight up fraud. So I said, No, I want to do it through the bank. And then the bank asked me, do you think it's fraud? And I said yes. And then it was like a do not pass go. It was call this number immediately. And so the friction just went way down. And it was, you know, a phone number that got me to a person right away. I'm not kidding. It was a nine dollar charge. So they were not talking about anything that's going to bankrupt Curious Squid anytime soon, but still counts. It still counts because you know there's blood in the water, basically. I found that interaction completely fascinating. My mood was like grumpy, like why are you making me jump through these hoops? To oh, you were dealing with this quickly and yes, and right away.

Rachel 55:31

Yeah, I think it's cool. These two lenses actually interact with each other quite a bit. Because I think that act of recognizing a vulnerable situation and then triaging with the appropriate amount of friction, because sometimes you have to slow down to speed up. You have to get the necessary information in order to take the correct next action. So these two lenses kind of uh speak to each other quite a bit.

Dan 55:55

Yeah. You've been listening to Unchecked with Dan Brown and Rachel Price. We were so glad you could join us. I hope you join us again. In the meantime, if you got some ideas for folks we can talk to, we want to hear about it. And if you find yourself using any of these lenses in your work, we also want to hear about that. Please do drop us a line, and we'll see you at the next episode.

SPEAKER_00 56:23

Unchecked is a production of Curious Squid. Curious Squid helps organizations like yours untangle complex information architecture and user experience challenges. Visit us at curious-squid dot com.