David Lazer, computational researcher: ‘If Trump wins, it would be a victory for disinformation’
The political scientist, renowned for his research on social media, calls for investment in independent platforms to examine the impact of big technology on society
David Lazer, a 58-year-old professor from California, delivered a packed lecture at the Polytechnic University of Catalonia, that focused on the need for a “transatlantic coalition” to study the influence of internet giants.
In a subsequent interview at the Barcelona Computing Center, the distinguished researcher at Northeastern University — who integrates computing and political science — discussed the significant challenges posed by a platform-dominated world characterized by misinformation and polarization
Question. How does disinformation alter the current democratic landscape?
Answer. If the citizen is not well informed, they are not going to make decisions that are aligned with their preferences. Disinformation and being misinformed fundamentally undermine democracy. There is a lot more disinformation floating around than there used to be, to be sure. But there are also a lot more people trying to correct misinformation. The interesting question that we don't really know is: what is the end result?
Q. What mechanisms do fake news use to reach and divide public opinion?
A. There are new entrants that don’t have the same traditional editorial standards that have entered the news ecosystem. And there’s also been the possibility for what is sometimes called the democratization of information sharing, where people select what they want to share and others see that content. Also, sometimes people express opinions, even if they only see part of the puzzle, and they comment. This makes it possible to get a distorted view of what is happening in politics or in the world in general. That is the contemporary concern. Clearly, we can see a lot more misinformation now, but people were also quite misinformed a generation ago. Are things worse now than they were before? Are they better in some ways and worse in others? It is probably a mix of both.
Q. And the misinformation is being spread on Twitter (now X), Facebook, Instagram…
A. There are a lot of mechanisms for it to spread. In social media, it’s small communities that share misinformation. On Twitter, for example, it’s a fairly small group that spreads a lot of misinformation, which wasn’t there before. But the consequences are greater when political elites or the media get it wrong. One of the biggest crises in the U.S. is opioids. And a lot of that was based on misinformation about how addictive OxyContin and other opioids were, which was spread through doctors and the media. We focus a lot on misinformation and terrible content online, but the smallest amount of misinformation we see from credible sources can be very consequential.
Q. Is mainstream media adopting disinformation more damaging than outright fake news?
A. Disinformation is accidentally amplified by the media or politicians. We can see, for example, that in the U.S. some politicians routinely say things that are not true, and then that is reported by the media. The question is how should the media report on that? Do you put a headline saying that such-and-such person falsely claimed something? But then the media outlet is accused of taking sides by fact-checking a politician in the headline. But if you don’t do that, some people don’t even get past the headline. The media gets co-opted by the spreaders of misinformation.
Q. How can we promote an information ecosystem that favors the truth?
A. We see actors trying to purposely manipulate the system by flooding it with with misinformation, but it is limited. The problem is not just to avoid disinformation, what we need as societies is to find a way to continue supporting the production of high-quality information that is independent, because disinformation is much cheaper. That is difficult, because many business models that existed to generate quality information have been obliterated, especially local and small media, which creates blind spots for our political and civic systems.
Q. To what extent will polarization influence the upcoming U.S. elections?
A. It’s clear that political polarization plays a huge role in what we see in the U.S. Each political side is less receptive to negative information about its candidate or to positive information about the other candidate. That makes the electorate very rigid and impervious to information, whether high or low quality. And why do we have this degree of polarization? It’s a very complex concept, it’s part of the dysfunction of the moment. Why does it happen? What role do social media play? Does the internet somehow contribute to that?
Q. Do you think social media is key in this moment?
A. People have always been social… Nowadays, the word “talk” includes things like WhatsApp or other types of instant messaging. My first experience with misinformation was when I spoke to a camp counselor who told me that John F. Kennedy wasn’t killed. I went home and told my parents and they told me not to believe everything I heard… We have to be careful about who we blame for misinformation… In the U.S., many Republicans believe that the 2020 election was stolen and that Trump actually won. And that’s because Republican political leaders are advocating that viewpoint. Politics has always been about creating information systems that make voters believe things that may not be entirely true. And there are very dark sides to this. If you believe that the election is being stolen, it can be the basis for strong reactions, like those we saw on January 6, 2021, when a mob invaded the Capitol. The question is where the U.S. is going politically, and not just in a few years, but next month.
Q. Could the storming of the Capitol have happened without social media?
A. There have been mobs that have overrun spaces before. I think social media created a space for dissemination and coordination that made it easier to do so, social media made it possible rally people together more quickly. That would have been harder in the past, but not impossible.
Q. Does Elon Musk’s ownership of Twitter benefit Trump?
A. It certainly matters, but it also matters a lot that Fox News has been very supportive of Trump. If I were Trump and I had to give one of the two up, I would give up Twitter, and keep Fox News. I think Twitter’s influence has been diminished by Musk’s partisan nature. It’s still a space for debate, but it’s been reduced, because it’s a clearly more unbalanced and toxic place. There’s some irony in the fact that Musk has perhaps undercut Twitter’s power.
Q. If Trump wins the election, could it be described as a victory for those who promote disinformation, fake news?
A. To the extent that Trump himself has been a key source of repeated disinformation. Yes, I think that’s true. He has promoted the myth of the falsehood of the stolen 2020 election, and so on.
Q. What can we do to reduce misinformation?
A. It would be good to create transparency about what the platforms are doing. And a lot of the capacity to study this is in academia. Academics should step forward and work with the media to create more transparency about what big tech companies are doing. And that would help determine the types of regulatory interventions that are needed and that, for example, in Amazon’s product recommendations, only scientifically validated claims are accepted.
Q. In one of your studies, you came to the conclusion that the algorithm does not increase polarization. One would say the opposite…
A. What we found is that if you change Facebook's algorithm, you spend less time on Facebook. If you then go to X from there, I'm not so sure that that decreases the polarization of your news consumption… In another article about Google, we looked at the filter bubble, to see if the algorithms are basically showing you the things you want to see. We showed that Google shows you pretty diverse content, but then people choose things that align with their preferences. The goal of the contemporary information ecosystem is to make it easy to access the information you want to consume. There's an active debate in the literature, which says that people are actually exposed to very diverse content online, but other arguments suggest that less diversity is consumed. If we think socially, we're generally disproportionately friends with people who are like us politically. If we look at partisan news consumption, we found that what Republicans read in The New York Times is very different from what Democrats read. And this has nothing to do with algorithms, but with the social processes of content dissemination.
Q. It’s all about us and our own choices.
A. There’s a plausible argument that we’re creating curated feeds that create politically aligned information flows. That said, there are a lot of other things going on in societies like the U.S. The likelihood that your neighbor is in the same party as you is much higher than it was a generation ago because geographically we’ve become much more polarized along political patterns like city versus country, the urban-rural divide. I don’t know how universal this is, but it just wasn’t the case 30 years ago.
Q. Do you think AI can help control misinformation or do just the opposite?
A. It’s a double-edged sword, we’ll see which way it goes, probably both ways. On the one hand, it makes it much easier to produce large amounts of false information, which has always been cheaper than producing valid information. It will allow you to produce images or videos that could be very persuasive. But it may also be better at detecting misinformation. We can also think about how it will affect search, the central way we get from one place to another. For example, if I have a question about facts, I could just open my phone and talk to it and voila, an answer appears. The interesting question is about whether our AI systems actually produce correct information or just plausible information. Currently, AI tools are quite oblivious to truth values. Will this always be the case? Not necessarily. A lot of it depends on how much companies value truth versus how central that is to their business. We’ve seen plenty of examples where the value placed on truth is very low. They value participation, retaining users, selling products, etc. more. And that is a big concern.
Q. There is controversial AI regulation in Europe. What is your opinion on it?
A. I don’t have a clear-cut opinion. AI needs to be transparent and meaningful and have analytical capabilities. We need to think more about how we create transparency similar to that applied to government economic statistics. How many internet statistics, how much money is invested in internet statistics agencies for the government? They don’t exist. We need to rethink how we create information and knowledge about these areas of enormous importance for contemporary society. And until we do that, rhetoric is going to vastly outweigh our knowledge. We need to think about how we invest in some kind of independent platform for creating knowledge.
Q. Finally, who do you think will win the US elections?
A. Do you have a coin that can be flipped?
Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition
Tu suscripción se está usando en otro dispositivo
¿Quieres añadir otro usuario a tu suscripción?
Si continúas leyendo en este dispositivo, no se podrá leer en el otro.
FlechaTu suscripción se está usando en otro dispositivo y solo puedes acceder a EL PAÍS desde un dispositivo a la vez.
Si quieres compartir tu cuenta, cambia tu suscripción a la modalidad Premium, así podrás añadir otro usuario. Cada uno accederá con su propia cuenta de email, lo que os permitirá personalizar vuestra experiencia en EL PAÍS.
En el caso de no saber quién está usando tu cuenta, te recomendamos cambiar tu contraseña aquí.
Si decides continuar compartiendo tu cuenta, este mensaje se mostrará en tu dispositivo y en el de la otra persona que está usando tu cuenta de forma indefinida, afectando a tu experiencia de lectura. Puedes consultar aquí los términos y condiciones de la suscripción digital.