Loading summary
Transcript68 lines
- [00:03]
Kurt Nickish
You're listening to Is Business Broken, A podcast from the Mehrotra Institute for Business, Markets and Society at Boston University Questrom School of Business. I'm Kurt Nickish. These next few episodes, we'll be diving into one of the more vexing challenges of our Internet age. The spread of misinformation on social media. From political lies to viral conspiracy theories, misinformation has reshaped our digital landscape, creating confusion, influencing public perception, and altering national debates. Today we discuss just how widespread misinformation has become, the roles that social media platforms play, and some potential solutions. Our guests today are Marshall Van Alstine, the Allen and Kelly Questrom professor in Information Systems at Bio Boston University Questrom School of Business, and Gordon Pentecoek, Associate professor of Psychology and Hyman Brown Faculty Fellow at Cornell University. Marshall, thanks for being here.
- [01:09]
Marshall Van Alstine
Good. Thanks for having us.
- [01:10]
Kurt Nickish
And Gordon, thanks to you too.
- [01:12]
Gordon Pennycook
My great pleasure.
- [01:14]
Kurt Nickish
So we're talking about how to manage disinformation in the digital age. To start, just how pervasive is disinformation today? In a modern context, it's very hard.
- [01:26]
Gordon Pennycook
To know how prevalent misinformation is. The problem is information is prevalent, and so we live in an information age. And so with that comes a lot of also misinformation. It also happens to be the case, at least in the American context. We see a kind of a breakdown of the quality of political communication. And so there's more outright falsehood than we would expect in political discourse, I.
- [01:49]
Marshall Van Alstine
Think there's also an interesting report put out by the Rand Corporation on what they call truth decay. In an environment of information overload, folks tend to fall back on tribalism and intuition for validation rather than taking the time to look things up just because there's too much. So the information overload, even if the misinformation isn't too much, it's still enough that it's causing decision error, if you will, and causing folks to make poor choices.
- [02:16]
Kurt Nickish
Gotcha. So just as we're on an information superhighway, we're also on a disinformation superhighway. Kind of disinformed or misinformed to death just because of the technology or there's a little bit more going on?
- [02:32]
Marshall Van Alstine
Well, if you look at historically, there have been lies even in Babylonian stone frescoes. So those have been around for a really, really long time. So is there something that's really different about today? One of the things I do think is different, that since social media have emerged, everyone's a producer where it used to Be certain vetted media channels such as New York Times or Fox or others with journalists here, everyone can be a producer of information so that the production has gotten so spread, so decentralized, it's harder to know necessarily what the valid sources are. And the information pollution problem, the infodemic, if you will, is much larger scope and scale than it had been previously.
- [03:15]
Kurt Nickish
Gordon, you've researched how people fall for misinformation. If that's the right verb. How does that psychology play out? Are there cognitive biases that make us particularly vulnerable, like what's going on there?
- [03:30]
Gordon Pennycook
So cognitive bias is a major role. Like very simple things can have pretty big impacts. For example, one of the first papers that we published on the topic was looking at the effect of simply repeating falsehoods. A single prior exposure to a fake news headline increases later belief in that headline. So even like implausible sounding things can be seen more plausible if they're repeated enough. And that's a straightforward just cognitive bias. In many cases it makes sense to trust things that are familiar, but when it comes to a kind of problematic information environment, that is maladaptive. And so that's like one major consequence.
- [04:07]
Kurt Nickish
And just to put a point on that, so I might hear something and think, well, that doesn't sound right, but if I see it about 17 times in my feed after that I will just automatically start to accept it.
- [04:21]
Gordon Pennycook
Yeah, I mean, it has an influence on how plausible it seems. Now the problem is that in many cases, the first time you see it, you don't even go down the step of that doesn't sound right because that already is assuming a kind of more deliberative take on people's social media feeds than is already happening. Another way to describe it would be people kind of engage reflexively. They go on social media often to kind of distract themselves from having to think about things. And so just they're kind of like passively consuming media. But that stuff is still having impact on their cognition. It's changing what seem makes sense to them, what seems plausible to them. And so this is like the shifting overton window. You know, the more you're engaged to falsehoods, the more they don't seem that false. They sound like something that you've heard before. Then you end up with beliefs that might surprise you later on.
- [05:08]
Kurt Nickish
Okay, what else? What else is happening?
- [05:11]
Gordon Pennycook
So one element of what I was talking about there is this being reflexive that is intuitive. You kind of automatically just accepting things that you're seeing without really thinking that much about them. And so we have research, for example, on how people go about sharing misinformation. And a lot of the time when people discuss this topic, they kind of make the assumption that when somebody shares misinformation, they're sort of doing it on purpose. They're like deliberately sowing discord. They're being politically polarized, they're trying to defend their ideologies, whatever. But what we find is something different. We find that a lot of the time people share misinformation because they haven't even considered whether it was true or false. They're kind of engaging in an automatic mode. Again, it comes down to kind of being inattentive to accuracy and to not spend enough time kind of thinking about whether something actually is true or false. And certainly people are not engaging in the extra step of fact checking or researching sources and so on.
- [06:02]
Kurt Nickish
I mean, is there an example of this? I'm wondering about. I don't know. If I read a story about unions asking for a better contract, and I just think, well, they're. I'm inclined to think biased, to think that they are. They're just trying to get a better deal, but they're not actually, you know, hard workers. If I see a story like that, I just might be more likely to spread it without really reading it, investigating it. It just confirms the bias I have.
- [06:27]
Gordon Pennycook
Exactly. And you might not even think that it has the potential to be false. You just think, oh, that's important, or this is something that makes sense.
- [06:36]
Kurt Nickish
Right.
- [06:36]
Gordon Pennycook
I'm going to pass it on. And then this happens even to people who do research on the topic. I have had the impulse, I've now kind of trained myself, to never kind of reflexively share anything. But you see something that in many cases might be outrageous or related to moral issues, and then you have that impulse to spread it, and you have to stop and just be like, wait, I should make sure that this is accurate and not just a small clip of a larger issue and a missing context or whatever.
- [07:04]
Marshall Van Alstine
So there's interesting data on that. I think. There's a study a couple of years ago that said that falsity spreads faster, farther, deeper, broader than truth in almost all categories. There was a study of Twitter a while back, and it's interesting if you think falsity tends to have higher levels of novelty and folks are sharing novelty, it's surprising, whereas the truth perhaps is less surprising. So there's another element where I think there's a certain amount of tribalism where sharing a particular claim is the party line. It's almost like wearing the badge of your sports team. I think it was a wonderful observation by Yuval Nova Hariri where he talks about, you know, it's easy to agree with the king if he says, okay, the sun rises in the east, but it's really, you're showing your member of the tribe if you agree the sun rises in the west. Right. Which is a completely ridiculous claim. But you're showing that you're a member of that group by doing it. It may be one of the other things that we're seeing also in such things as the insurrection or the stop to steal or that the election was stolen. Again, you are declaring yourself a member of that group.
- [08:03]
Kurt Nickish
So how do the platforms, these social media companies, these tech companies, handle this? On the one hand, they like the fact that people are sharing things. They're active on the platform, they're spending time on the platform, they're engaged. That's something that they want to foster and facilitate and make easy. On the other hand, they also want there to be a good experience for the people who are getting that shared to or shared with. Right. They want them to get good information, useful information. You know, the same way that a phone book doesn't make sense if you can't really trust all the numbers in it. If 20% of the numbers are wrong, it's not a good experience to use a phone book. Is this a fair way to simplify and exaggerate the conundrum that's facing social media companies right now?
- [08:59]
Marshall Van Alstine
Well, the business model is one that's often described as if it's enraging, it's engaging. And if they can keep you engaged, they can sell advertisements. So it's one in which the business model actually tries to get folks to participate. This is what we heard with Francis Haugen's testimony before Congress, putting profits over people. That's what Sushana Zuboff identified in surveillance capitalism as another element of trying to observe everything that you can do in order to model you and then sell you in your weakest moment when you're most desirous of some particular product.
- [09:29]
Kurt Nickish
At the same time, though, a lot of people talk about some of these platforms as cesspools in places where they don't feel comfortable sharing opinions because they'll get screamed down. It seems like there's a lot of consumer pressure on these platforms to change their experience.
- [09:45]
Marshall Van Alstine
Well, you make a good point. I think even the folks at Techtor argue that in many cases, the public pressure does try to get the platforms to reform Some of their behavior to make it a better experience. Apple curates their ecosystem to remove malware, to remove pornography, to remove hate speech and things of that sort. So different platforms do make different choices. The one caveat that I would make is that in many cases they'll try to create a better experience on platform even as they shed a lot of the responsibility for damages that occur off platform. So they don't experience the insurrections, they don't experience the rioting in the uk, they don't experience judicial interference. So engagement where they benefit is something where they don't experience the cost. So some of the things they will manage on the platform, but I'd argue they don't manage the ones off platform.
- [10:38]
Kurt Nickish
Okay, so what are some examples? This is open to both of you. What are some examples of platforms successfully combating misinformation?
- [10:51]
Gordon Pennycook
That depends. That's a good question, I think. Okay, well, I think there is one clear good example of a successful policy that was instated and that was it used to be called Bird Watch, but now it's called Community Notes at Twitter and that is crowdsourced fact checking. The Community Notes is actually used at a reasonable frequency. People do find them useful and they are quite good. So that's one policy. But of course it hasn't fixed Twitter or now X if you've gone on it has not like solved the problem of there being accessible of information. It's just. But it is a successful policy. It comes back to the. The overarching issue in my view is that social media companies aren't committed to understanding the issue, the problem. In fact, their incentive is to not understand it, to try to do things that sound like they would work when rolled out so they don't get backlash. And it makes it seem like they're doing something, but that technically don't maybe work because you don't really want to damper engagement on your platform. And so they don't put sufficient resources into actually understanding what's happening and like how misinformation spreads and collaborating with researchers and that's I think, the underlying problem.
- [11:59]
Kurt Nickish
I mean, you say that, but they spend tons and tons of money. I mean, maybe X is an example of where the company is kind of crowdsourcing that through community notes, but YouTube, other places, I mean they, it's an expensive cost for them.
- [12:14]
Gordon Pennycook
Yeah, but that's not understanding the problem. That's geared at trying to improve the platform. If you, if you really want to find a way to improve the people, the way that people engage with the Platform, you have to understand the underlying kind of psychology of what's happening on the platform. And in fact, if anything there's a strong kind of like distaste for doing experiments. And some of this is public backlash. They don't want to do things. And there are some cases there are interventions that could be rolled out that are informed by psychological mechanisms, but that seem to heavy handed, for example, and so they wouldn't want to do that because then people would not like it. And therefore that undermines the business model.
- [12:54]
Kurt Nickish
We'll talk in a second about how this should change the possibilities to improve the space. But to that counterexample, are there some examples of platforms mishandling?
- [13:05]
Gordon Pennycook
I'll give an example that relates to what I said, which was so after the 2016 election, when fake news became a giant story, what Facebook did, the very first thing that they did was they added this label called disputed by third party fact checkers. It was like a little thing that just tried to tell people that hey, this is something that third party fact checkers think is maybe false. And they had not apparently tested it in any kind of meaningful way. They just kind of, it was something that was proposed as something that internally would make sense and then they just ran with that and millions of people saw the thing. We did some research showing that like, you know, if there's a lot of these labels and some people are going to think that things that are unlabeled are more likely to be true. And so it really depends on how it's rolled out that you can, you have to test these things to see what actually works. And so they stop using it. But then they have other ways of doing similar things. For a company that has access to such a huge amount of behavioral data, they don't take advantage of that. There should be testing to see what the most effective interventions are, but they're really just using interventions that make sense intuitively to people so that it deals with the PR problem and not the underlying problem.
- [14:12]
Kurt Nickish
Yeah, that might be a case of moving fast and breaking things, but it is giving me an appreciation for how complex these systems are.
- [14:19]
Marshall Van Alstine
So I want to give Gord real credit for identifying the implied truth effect as one kind of intervention. What's really interesting is the kind of research that Gord does and Brendan Deyhan and others do is it would be great if you have access to the data. But Facebook and Twitter and TikTok and others have shut off the APIs. They've turned it made it impossible for academics. They have created Some databases, and of course these are curated databases. We don't know what's actually happening in the background. If we had legitimate access, then it would be easier to run AB tests, controlled experiments to gather data and see what would really be effective, to see if it would work. But they have made that increasingly difficult to do. There was a study out of nyu, for example, where they'd gotten permission to install browser plugins to track what misinformation or ads are being shown on Facebook. Facebook said, no, no, no, that's the advertiser data. Well, that's bunk. Clearly these are the things that users themselves are being shown. And so they ought to have the right to be able to monitor what it is that they're being shown. So I think in some sense we come back to it, but I think user rights would help. But the platforms themselves have made it difficult for legitimate academic research to dig deeper into the problems.
- [15:29]
Kurt Nickish
Gordon, is that the case for you too? You had an article in Nature this month about social media and sort of the difficulty of the perception problem of moderating content.
- [15:42]
Gordon Pennycook
Yes. So that is a problem for us. In fact, that paper uses data that we would no longer have access to because Twitter does not have an open API anymore. In that paper, what we show is that in fact it's funny because the paper defends the decisions of the social media company, in this case Twitter at the time. There's a bunch of outrage about they were censoring conservatives. And you do find that in the aftermath of the 2020 election, the Conservatives are more likely to be booted off of the platform. But this is explained more by their sharing of misinformation than it is by whether they're conservative or liberal.
- [16:16]
Kurt Nickish
Yeah. You found even pro Republican groups looked at the policies and thought that they were fair.
- [16:21]
Gordon Pennycook
Yeah. I mean, that is, if you define misinformation based on politically balanced layperson ratings of the quality of sources, you still see that that value, that like rating of misinformation is a better predictor of being suspended than being Republican, which means it's about the behavior and not the ideology, which again defends the platform. But now we can't do these types of audit studies or investigations because we don't have access to the data.
- [16:46]
Kurt Nickish
So that's really fascinating. And it also gives you a little bit of empathy for the companies because they realize that even if they do something fair, they're going to be seen as sort of being heavy handed one way or the other in how they're going about it. Where should innovation in this space come from to come up with better ways to handle misinformation from the platforms themselves, from policy, from competition, emerging companies, I guess, where's the most promising fertile ground for coming up with productive changes?
- [17:22]
Marshall Van Alstine
So I love the question. I would argue that platforms themselves alone should not be able to do that. You're asking the fox to guard the hen house in that case, to fix their own business models in a perfect world, police themselves. Exactly. One would hope they would do better. And actually there's some really wonderful research on self governance. And self governance tends to work if there's sufficient pressures or there's even meta governance. But I would think that good sources would be neutral parties, such as the kind of research that Gord has just articulated from competitors, if they actually had access to data to show when misdeeds were being done. And also if there were regulatory access to be able to monitor. If there are transparency, users could then defend their own interests. But at the moment they don't have the transparency, don't have the access. So those are at least three alternate sources in addition to, and certainly beyond the platforms themselves. And again, I think we do really need to go past self governance because I think we need alternate perspectives in order to get legitimate answers.
- [18:21]
Kurt Nickish
Yeah. Gordon, anything to add?
- [18:24]
Gordon Pennycook
No, I agree. I mean the social media companies have had plenty of time to prove to us that they were going to deal with this in a creative and innovative way. And I don't see that being the answer.
- [18:36]
Kurt Nickish
Interesting. And so just to underline this a little bit more, what do you recommend or what would you like to see?
- [18:44]
Gordon Pennycook
I guess my suggestion was to let me do it. I'm just joking. I mean policy. We need something that changes the levers so that there's forced innovation, so that we can come up with evidence based solutions and not just try things that sound like they make sense.
- [18:58]
Marshall Van Alstine
There are one or two solutions we can propose, so these are some things that we in fact have worked on. One thing I adamantly propose is one called an in situ data. Right. Basically the idea is you can bring the algorithm of your own choice to your own data wherever it's resident. Imagine that were in place at the time the New Yorkers NYU was running its study on Facebook. Then users themselves would have chosen to allow this other research in to be able to see. Then people like Gord and Dave Rand could be able to do great research and you'd be able to monitor. I'll give you another good example. If you had it in C2 data right then the Drivers on Uber could actually see, they could import an algorithm, see what's being shown to the other drivers and they could then on collective action basis, retake some of their power. Or imagine you had an in situ data right to determine which filter you're being shown on Facebook or Instagram or TikTok. If that's the case, you might then choose a BBC filter or a Breitbart filter or a Google filter on there. The benefit of that is why would you do that? You might get then the choice of the information that you want, but you might also get a share of the ad revenues, in which case Facebook is like, well heck, we don't want you to take the Google filter, we'll share some of our ad revenues. What that does then it then means you become the source of the bottleneck. So not only do you regain the transparency and some of the power, you regain some of the value created from that. So that's another mechanism we might be able to use in order to actually rebalance the scales, share the value and create transparency and then create innovative research solutions as well.
- [20:34]
Kurt Nickish
That's a really interesting technological approach. What about AI? How could AI be used to combat disinformation and misinformation? Gordon, you have done some research on this.
- [20:46]
Gordon Pennycook
What we did in our experiment is we had people have evidence based conversations with an AI. In this case it was GPT4 Turbo about conspiracy theories. The use case here is essentially pretty straightforward. And so what we do in those experiments is we have people have a conversation about a conspiracy that they believe in in their own words. And the AI is being told to basically persuade them not to believe in these things that are dubious. And it's using evidence based counter arguments. Just like someone might talk about 9, 11 being an inside job. And it says something about like the reason I believe this is because World Trade Center 7 burned down but wasn't hit by a plane. And then AI would draw on reports saying, well, that the reason that it burnt down was because debris hit it and that lit on fire and so on and so on. So what we find is like an 8 and a half minute conversation after a few back and forths with the AI. That's on these evidence based conversations, people who believe conspiracies decrease their belief by 20%. Fully a quarter of them changed their mind. They stopped believing in the conspiracy after the conversation and it was only about eight minutes long. So that just shows the power of having these kinds of evidence based conversations with artificial intelligence.
- [21:54]
Kurt Nickish
Right, that's really interesting. And Maybe easier than producing lots of videos or getting people to read counter arguments or go to snopes.com or whatever. It also seems like that technology could be used the other way around.
- [22:07]
Gordon Pennycook
It could. We didn't of course, test that because we were trying to do more ethical things, although it was the case though that in this sample, like a small percentage of the conversations were about true verified conspiracies like MK Ultra, for example, and the AI did not dissuade people from believing things that were true then. The reason, I think, is that it's going to be more difficult to come up with a good argument if you don't have evidence on your side. You could in theory build an AI that's really good at making up evidence. That is a more difficult problem than just drawing on sources that are good and parroting what's already been written. But it is potential, something that we might see down the road and it is a worry.
- [22:47]
Kurt Nickish
I mean, we're talking about misinformation where people are passing things along because of the psychology of it. But there are industries out there where people make money trying to show that the world is flat and getting the advertising views from people looking at the videos about that. So that may be just a very small part of this problem, but it is a market right there. And I just wonder what's the role of markets in this problem and its potential solution?
- [23:18]
Marshall Van Alstine
Well, there are a couple of different approaches that you may consider in that Google, for example, has tried to demonetize or actually make it harder to put misinformation ads on vaccines during a pandemic. I think there are some other elements on blatantly false information. They won't run that ad and they won't run a legitimate ad on a website that actually has that kind of content. So I think they're trying to address some of those issues in some of the most extreme cases. But you raise a different question, which is that in most cases it's concentrated market power. The Google is making decisions, or Amazon is making decisions, or Elon Musk is making these decisions. They're not market based. And in fact, one of the things I think we really do need to do is to try to design market mechanisms to decentralize these choices, to take the central powers out because of the power for mass persuasion, because of the power for doing illustration on that basis. I think we need more decentralized, more distributed mechanisms for validation and not those that are centralized.
- [24:23]
Kurt Nickish
Your recommendation or your idea to have certain filters and give power back to the users. That's an example of something that may be a market solution, but also not something that the current companies with the market power want to give up. I want to ask where we draw the line between free speech and misinformation, but it's not necessarily an either or choice or balance of those two things. But we are talking about trying to massage control an information system at a time when you know one of the strengths of the Internet and the platforms is that it feels like a huge public square. It is a place where you should be able to exercise free speech. How do you see the tension there?
- [25:10]
Gordon Pennycook
One thing I would clarify there is that it already is being selected. That is like there is no platform that just randomizes or shows you algorithms are involved in all of it. So it orderly is kind of inherently sensorial. There's like something that determines what you're going to see. The question is what should we use to do that? You can use a version like TikTok for example is a kind of model where it's just mostly how much time you spend looking at something it seems and whether you like something, all that kind of stuff. It's usually kind of just user feedback going into the thing. But there still is like certain things go viral and they're pushing their thumb on the lever in certain ways. They also don't want to show people computer stuff and they might downrank that kind of stuff. Every platform has some element of that. The question is, what are we going to prioritize as a society? What do we think people ought to be seeing? What should we be amplifying?
- [26:07]
Marshall Van Alstine
Algorithms are already there effectively deciding what content we see.
- [26:10]
Kurt Nickish
It's almost like your filter that you're talking about. Instead of being able to choose the filter, the companies are basically personalizing a filter for you.
- [26:19]
Marshall Van Alstine
But it's their version they get to censor. What's interesting is if you look at it carefully, we've traded under the First Amendment, the absence of government censorship for the presence of corporate censorship. They get to decide they are trying to exercise what's in the law described as publishers rights. They have the editorial discretion over what you see. A fascinating example of this, if you turn the clock back about three months, Joe Biden had just stepped down and let Kamala Harris run for the presidency. At that time, the flood of people trying to sign up to follow the campaign was so large the bots were triggered and actually blocked followers. What's interesting about that is that that was a technological accident, but it could be the actual editorial choice of Elon Musk. In fact, he could choose under publishers rights to separate Harris from her followers. That's wrong. That's violating freedom of assembly. That's violating freedom of speech. And it's a way in which corporate encroachment on those rights has gone past the point where in my view, that it's legitimate any longer. And that's what's feasible currently under section 230 of the law. And I think it's time we restored some balance in listeners rights. So the folks that you choose to listen to are the ones you get to listen to under the sets of filters that might actually work as opposed to simply under the shroud of corporate censorship, as opposed to government censorship.
- [27:45]
Kurt Nickish
So that's really interesting. Like we think of Twitter now, X as a platform that's supposed to be less politically correct, more open, no censorship. Right. That's the brand practically. X is censoring is what both of you are saying.
- [28:02]
Gordon Pennycook
Yeah. And I guess another way to think about it is like not what are you censoring, but what are you amplifying? That's one thing that has changed a lot on X since Musk took over is that different things are being amplified. And so that's an editorial decision. And people think about it in terms of like, oh, it's free speech, but it's really about who you're giving the megaphone to.
- [28:22]
Kurt Nickish
Yeah. You see why that can be a problem. What does an effective ecosystem look like for information distribution on digital platforms? One that's going to kind of balance truth, freedom of speech, business interests, which we also care about.
- [28:40]
Marshall Van Alstine
This is a much deeper question, and I actually think it's one where we need to design marketplaces where you could expand listeners rights, which I just articulated, but also expand speakers rights in an interesting way. So I pose the question, do you actually have the right to be heard? The First Amendment operates as a negative right on government and censorship, but it doesn't actually necessarily give you the right to be heard. Another version of this is a Silicon Valley expression. You have the right to speech, but not the right to reach. And if you talk about filter bubbles, how would you pierce a filter bubble? You'd want to be able to gain access to that filter bubble to be able to present information. And one of the things we could use, we could try to adapt a couple of ideas from Ronald Coase in creating true marketplaces that actually balance the rights of source and destination of buyers and suppliers, or in this case listeners and speakers. So we ought to be able to balance the rights of listeners not to hear things that they don't want, which would include pollution and harassment and other things that would be unpleasant, but then also the rights of speakers to be heard. And one of the ways that Coase tells us an excellent means of doing that would be okay. When does one right dominate another? It's on condition you accept responsibility for using it in the right way. You accept responsibility for legitimate use of your rights. So if you're going to not accept that responsibility, then you shouldn't be able to override someone else's right. In this case, if someone doesn't want to hear what you're saying, then if you're not willing to accept responsibility or vouchsafe what your claims are accurate or legal or not harassing, then they don't have to hear you. But if you will vouchsafe the validity, the legality that is non harassing, then you have the right to be heard and to pierce a filter bubble with the facts to present your views or different perspectives.
- [30:29]
Kurt Nickish
So that's really intriguing. Gordon, what about you? What's your sort of vision or hope for an effective ecosystem for this?
- [30:37]
Gordon Pennycook
I'd say this is beyond my pay grade.
- [30:39]
Kurt Nickish
Okay. Yeah, yeah, right.
- [30:42]
Gordon Pennycook
I mean, you have to obviously balance the preferences of users with some sort of determination of quality, but there's no simple solution to that and no simple way to determine what ought to be listened to and what not ought to be listened to.
- [30:55]
Kurt Nickish
Yeah, and we need more research to understand the problem.
- [30:58]
Gordon Pennycook
That's. That's my only song I sing. Hey, guess what? We need more research. It's just weapons that I do research, so it's a little bit biased, but that's what my perspective is.
- [31:06]
Kurt Nickish
Yeah. Gordon, thanks so much for sharing your voice and expertise. Really appreciate it. Pleasure, Marshall. Thanks to you too.
- [31:14]
Gordon Pennycook
See you, Marshall.
- [31:15]
Marshall Van Alstine
Take care. See ya. Bye bye.
- [31:20]
Kurt Nickish
That's Professors Marshall Van Alstine and Gordon Pennycook. Next week, we dive further into how to regulate platforms. In an age of fake news, how do we reconcile the protection of free speech with the need to prevent harmful misinformation from spreading online? Is it even possible to strike a balance? That's next week. To get that episode and more, please follow the show on Apple podcasts, Spotify or wherever you listen. Thanks for listening to Is Business Broken? I'm Kurt Nickish.