Internet Trolls in the U.S. and Mexico, with Saiph Savage

DEVIN STEWART: Hi, I'm Devin Stewart here at Carnegie Council in New York City, and today I'm speaking with Saiph Savage. Professor Savage is assistant professor at West Virginia University, and she is also affiliated with Carnegie Mellon University as a visiting professor.

Saiph, great to see you today.

SAIPH SAVAGE: Thank you for having me. I'm very excited to be here.

DEVIN STEWART: You're on the cutting edge of a lot of different areas.

SAIPH SAVAGE: Thank you.

DEVIN STEWART: Information technology. This is also a part of our ongoing podcast series that we're calling Information Warfare.

You design interactive tools that empower communities to coordinate volunteers to produce collective action and trigger real-world change. So you're an activist scholar and also a technology expert.

Can you give our listeners a sense of your work and your approach and how you are approaching what we call "information warfare," but you could call it "disinformation" or "propaganda" or "fake news"? What is your unique angle?

SAIPH SAVAGE: I like that you mentioned that I'm an activist. I take these problems from a computer science perspective on one hand because that is my training, and then I also do have an activist background as well, which I also use to design my systems.

On one hand, within my research lab what we're doing is we're using large-scale data to first start to understand what kind of warfare is happening in the wild. How are people being targeted? For instance, we recently did a large-scale study about how Latinos were being targeted for the midterm elections. Then, once we have a better understanding about what exactly is going on out there—how are they maybe being targeted on Reddit, on Twitter, how are they being attacked, what type of fake news are they being exposed to, what kind of political trolls exist—we take a step back and then we start to say, "Well, what did we learn from all of this, and what types of tools can we create to start to help people to address some of the problems they are experiencing?"

My research focuses on one hand on doing large-scale data analysis, and here I use a lot of computational techniques, but I also use qualitative studies. I do pick up some things that ethnographers do to understand how are people interacting with the data and what is going on with them. We also do a lot of interviews and surveys.

Once we have a deeper understanding, we build systems. I also like to build systems where we're coordinating citizens to empower them to solve the different problems that they might have, some of them, for instance, fighting disinformation.

DEVIN STEWART: Since we just had the midterm election and you mentioned your study of the targeting of Latinos, what did you find?

SAIPH SAVAGE: That's very interesting. We were studying how Latinos were targeted on Twitter and Reddit. Who exactly was mentioning Latino-related content that was around also the midterm elections?

DEVIN STEWART: These are American Latinos, yes?

SAIPH SAVAGE: Yes, U.S. Latinos.

We studied within Reddit what types of posts were being created around Latinos in the midterm elections, and on Twitter also who was mentioning the Latinos and the midterm elections, so these two elements.

What we first started to uncover was that on Reddit the ones who were the most active were actually the communities that previous work has identified as political trolls. What is interesting is that these political trolls had really sophisticated techniques for engaging audiences around politics and around Latinos.

The political trolls would create spaces where they would invite, for instance, candidates who were also Latinos, let's say maybe some of them running for governor or they wanted a seat in the Senate. They would invite them into a subreddit to start to participate. They would create what are called "megathreads," where they would have these long, detailed discussions with others about the elections and how this impacted Latinos, what type of things candidates allegedly were doing to help Latinos, so they had these deep conversations. They would also post jokes. They would post jokes against immigrants.

What we found was that there was a lot of action from these political trolls, and all of them were mobilizing Latinos—and more of those Republican—and the ones that were more neutral that were pushing just to get Latinos to vote and maybe some of them were also a little bit more Democrat were actually much less active, and these were more non-governmental organizations (NGOs).

What we started to find was that there was somewhat of a gap where Latinos were not exposed to more neutral actors that were mobilizing them, but rather they had these political trolls and people more in the extreme who were engaging with them.

DEVIN STEWART: What were the jokes like? When they were invited to come onto Reddit, was it a kind of trap, or was it sincere?

SAIPH SAVAGE: That's a great question. It was sincere. The candidates that they tended to invite were candidates that they supported. A lot of the things that these political trolls handle online is that the media is portraying them under a different light. So it was also a space to give their candidates a way of presenting themselves without the alleged bias from the media.

DEVIN STEWART: So it's a friendly troll.

SAIPH SAVAGE: It's a political troll that supports Republicans, so toward Republicans they are friendly and they explain the ecosystem according to the Republican lens. With their candidates they're being friendly.

DEVIN STEWART: Their candidates are Republicans as well?

SAIPH SAVAGE: In all cases, they were. These political trolls in general identify themselves as pro-Trump trolls, so they've been supporting the president's agenda.

Also, within this, they would make jokes around illegal immigrants, let's say around also the caravan that was happening, and basically it was their support.

DEVIN STEWART: What was the character or tenor of these jokes? Did they sound hateful or angry? Were they actually lighthearted?

SAIPH SAVAGE: I would say that they touched things that were currently happening and then give them a humorous spin. For instance, they would say, "Well, if Hillary Clinton had been elected president, right now we would probably be accepting all of these illegal immigrants, and we would even be gifting them to Angela Merkel who just arrived for a visit. They would each be exchanging these refugees." That was the type of jokes that they would make with each other where they take a current political situation and find a humorous angle for them.

In some cases they would play with stereotypes. Another one was they posted a picture of a woman who was obese and a Hispanic who was marrying her. The title was, "When Your Visa is About to Expire by Midnight," so playing into some stereotypes that people have where a Hispanic is always going to need a green card to survive and is willing to do anything. They would play a lot of times into these stereotypes.

DEVIN STEWART: The point of the jokes, the objective, was it to goad people into engaging with them?

SAIPH SAVAGE: That's a great question. My impression is that the jokes are a little bit to create a type of culture. On one hand, it's this notion to get people to reflect maybe that Hillary Clinton would have just accepted any type of refugee, and it would have been like this excessive acceptance of any type of refugee, any type of illegal immigrant that wanted to enter the United States.

I feel the jokes are aimed at getting people to understand what is the political agenda that they are supporting. In this case, the political agenda that they're supporting is that we should not be continuing to accept illegal immigrants or refugees.

DEVIN STEWART: It's a kind of signalling?

SAIPH SAVAGE: Yes. That's a great term. I feel it's a type of signalling where they're promoting a type of, to get people to understand what is the political agenda that they are supporting.

I think that they also do it a lot because in certain cases they are opposed to what the mainstream media has been pushing. In order to get people who maybe have not been exposed to them to rapidly understand, they need to repeat it and repeat it and repeat it.

DEVIN STEWART: What are they opposed to that the mainstream media has been promoting?

SAIPH SAVAGE: One thing that I was right now thinking about was globalism. In general, the mainstream media has been promoting an agenda where globalism is seen as something that is good. These political trolls, we had a lot of content where they created a lot of jokes against globalism. One joke was where maybe Starbucks was hiring a lot of foreign Muslims while U.S. veterans were out on the streets asking for five cents to survive. They created a lot of jokes around those lines.

I think they especially were trying to promote that because they feel that the mainstream media has been pushing agenda A, and they're trying to get people to see that maybe agenda A is not the best for the country. So they constantly are posting jokes to get people to reflect.

DEVIN STEWART: How effective was the engagement of the candidates? Did the candidates take the bait? Did they interact? Do you have a sense of the outcome of this?

SAIPH SAVAGE: What we were finding was that those were some of the posts that got the most engagement from people.

DEVIN STEWART: Which posts?

SAIPH SAVAGE: The posts in which they would bring in the candidates. They would bring in a political candidate who would basically answer any questions that people had.

DEVIN STEWART: On Reddit.

SAIPH SAVAGE: On Reddit, yes. They had sessions which are called "Ask Me Anything." They brought in candidates, and the candidates would respond to any type of question that the citizens were asking them. This was some of the most popular content that was created within the site.

What we noticed was that on the other hand you had NGOs that were not pushing a particular political agenda, so they were not telling people, "Oh, vote Republican or Democrat," like the political trolls were doing, but they were just trying to get Latinos to go out and vote. They were receiving much less engagement from people on these social media platforms.

The political trolls, on the other hand, were very effective at having many different dynamics for engaging, basically making it even more motivating for people to get involved.

Another thing that we also noticed which we didn't see from the Democrat side was that these political trolls would also post pictures of Latinos who were supporting Republicans. They were also very good at showcasing a brand of, "Oh, you're a Latino, and it's actually cool to be supporting Republicans." They also managed it as a way of critical thinking because the mainstream media, they were promoting it as, "The mainstream media is telling you that you as a Latino should not support Republicans, but here is why you should," and they're branding it as, "You're a critical thinker for stepping outside what the mainstream media is trying to promote to you."

DEVIN STEWART: Interesting. It sounds like you're also referring to the generation of memes among Republican social media people. I think there's a conventional wisdom, or at least there's the assertion that the Republicans do a very good job or a better job of creating memes compared to the Democrats. Is that true? What do you make of that claim?

SAIPH SAVAGE: That's a great question. We actually interviewed a lot of people who were involved in what was called "The Great Meme War" in 2016. Basically, all of them were Republican. They had very sophisticated techniques of operating.

They created a type of A/B testing where they would create an image collectively, and then they would post it in certain Facebook groups, see how people reacted to it, and then say: "Oh, no. You know what? People interpreted this wrong. We need to tailor it." They were working collectively to identify what was the best way that they should present a given message.

Then they also had many different groups in which they tried to infiltrate themselves. Within Twitter, for instance, they also had networks where they would say: "Okay, these are the soccer moms. Hey, soccer moms, we need your help in distributing these images." So the soccer moms would engage with them. Basically, they were very effective in being able to get into different networks and learning what type of content they should post in these different networks.

DEVIN STEWART: Do you have any theories about why the right or the Republicans tend to be more businesslike about it with the A/B testing compared to the Democrats? At least, that's the assertion out there.

SAIPH SAVAGE: That's a great question. On one hand, from the interviews that we did I think it was also in terms of that these individuals felt that they were fighting the system. It was a lot about feeling that they were part of an anti-establishment movement, and so they basically dedicated a lot of their own personal time to be able to fight what they considered was the establishment, and the anti-establishment just happened to be under the Republicans. That was my impression. It was something they were literally willing to dedicate a lot of their time. They had complete skin in the game, so they dedicated time and resources. I know a lot of them invested a great amount of money into this. Some of them were working full-time on a lot of these campaigns.

But I think their motivation was a lot that they felt it was a fight against the establishment, and the anti-establishment candidate happened to be I think in this case Trump, so they backed him with the Republicans.

DEVIN STEWART: That's very interesting.

Looking at one of your papers here, your research paper is called "Mobilizing the Trump Train: Understanding Collective Action in a Political Troll Community." What else did you learn from that study?

SAIPH SAVAGE: In that study we were also studying on Reddit how were these political trolls, in particular the ones operating within a subreddit that is called "The Donald." The Donald is a subreddit that the media has called a "political trolling community," and we wanted to understand—they call them political trolls because they have organized a series of campaigns against different candidates that have opposed Trump. They have done online harassment, doxing. They've been organizing a lot of nefarious behavior. We wanted to understand how do these political trolls organize. How do they get people to participate in their cause? What type of calls to action do they make to citizens?

We collected basically a full year of their posts a little bit before Donald Trump announced his candidacy until a few months after he won the election. From those posts we wanted to study what types of calls to action did the political trolls make. How did they get people to get involved with them? There we used some clustering techniques with qualitative analysis, and we identified that political trolls had three main ways in which they called citizens to action. On one hand, they used what we called "troll slang," so they used a lot of nicknames for people, slang, troll slang, and they used that with memes as well to get people to take action.

Within this, for instance, they had what was called "Pepe the Frog," which was this green frog that they identified with, and they created different memes with it. They would call each other as well different nicknames, and based on that they would call people to action. They called each other centipedes, and they would say: "Hey, centipedes. We need you to re-tweet this to showcase that Hillary Clinton is a crook and Bill Clinton is a rapist." They created content where basically they insulted people that opposed Donald Trump and then used that to further engage people to take action.

We identified that that technique was actually not as effective. With that technique they weren't able to get a wide range of people to participate. But it was effective with the ones who had been long-term participating with them.

Another thing that they did was just have really quick direct ways in which they asked people to participate. For instance, they said: "Hey, guys. We need you to upload this picture of Donald Trump because he should be on the first page of Reddit." It was a clear call to action, and people participated with that.

That was more effective than the political troll style, and they were able to get people to rapidly make—this second technique was just being direct, and in general it focused on viralizing content. A lot of their calls to action were in making their content viral on Reddit so that other people could be exposed to it.

But what we identified was that the most effective way that they called people to action was by explaining the political ecosystem in detail to others. They created long posts where they explained, "Oh, so Hillary Clinton did this, this, this, this, this, and right now this and this and this is happening, and that is why we need you to do X." Those were the types of posts that would receive the most amount of comments, the most number of up votes. It seemed that explaining to citizens in detail what was going on was most engaging for the political trolls.

I think there it goes back to what we were discussing before, in which political trolls sense the media is opposing them, and the media might be presenting a different notion about what X or Y event means. They have to take the time to explain to people their point of view. That was the most effective for them. That was one of the biggest takeaways we found from this paper, which is political trolls explain in detail why other citizens should participate.

DEVIN STEWART: Another area of your work is in Mexico. You have been looking at fake news in Mexico and also how to fight it. What's with the fake news environment in Mexico? How long has it been there? We talked about this a little bit when we first met in Washington.

SAIPH SAVAGE: One thing that I want to mention is an interesting thing that we are identifying is that to fight fake news we've identified that activists are actually taking an approach that is somewhat similar to what political trolls in the United States are doing. They're explaining in detail to citizens what is going on, and that is effective for them to get more citizens to participate with them.

For instance, we recently were analyzing how misinformation was happening in the 2018 presidential election that just happened in Mexico. It turned out that some activists started creating content to explain to citizens misinformation that was going on.

What we identified was that they even created manuals, long manuals that explained to people: "Okay, you know what? The media and these opponents are saying X or Y about this presidential candidate and his ideas. This is not true. This is what is actually going on." Some of that material was the one that became the most viral. I think there is value in understanding that citizens are not necessarily dumb, and they want to understand what is going on in the political ecosystem. They do actually want to take the time to understand it and that is the content that they are engaging with. That for me is a very interesting finding, that both in the United States and in Mexico what was most effective, what got citizens' participation, was content that explained to them in detail what was going on.

Within Mexico we looked at the presidential election. We also recently looked at how citizens organized around misinformation in a big earthquake that happened in Mexico. Here we wanted also to understand how did citizens fight misinformation. How were they operating within different social media tools?

What was also interesting for us was understanding that you suddenly had these experts in technology creating tools, but creating tools wasn't what was important for them but rather creating a community around the tools so they could take action in case something else happened.

Basically, in Mexico we've been looking a little bit into these two areas about the presidential election and also within the earthquake. In Mexico you also have something that are called "pay trolls," these farms of people who were being paid allegedly by a certain political party to take small actions for trolling and harassing others.

DEVIN STEWART: The conventional wisdom in the United States is that trolls and bots and fake news played a big role in the 2016 election. What is the common view in Mexico about the 2018 election of Obrador? What's the common view there, and which side of the political spectrum did the political trolls align?

One last question: Does the conventional wisdom in Mexico match what you've found from your research?

SAIPH SAVAGE: Those are great questions.

DEVIN STEWART: Three big questions.

SAIPH SAVAGE: The first one I guess was what happened in these elections. I think one important thing to think about is in 2012 Mexico had allegedly a big bot network that supported the presidential candidate who won. That political party had allegedly created a bunch of bots that—and this is 2012—went out and were super-supportive of him. What this created from my perspective was that the Mexican citizen was very accustomed to bots and to seeing bots operate and also seeing trolls operate because since 2012 they just got invaded by them.

Recently, a colleague, Phil Howard from Oxford, did a study about how Mexicans were engaging with fake news for the presidential election. What he found was that interestingly Mexican citizens were hardly sharing fake news. They were sharing these more trustful sources. I think it goes a little bit back into that in 2012 we lived a very difficult situation where we were invaded by bots, and we had there our fake news crisis. But the citizens learned from that, and I think that empowered citizens to not necessarily be engaging with bots and even with hashtags.

I think what bots were doing in Mexico a lot was positioning certain hashtags in support of the political candidate who was paying for them. Citizens became very savvy in learning how to even play with those hashtags to start to push the narratives that they now wanted to push instead of the ones that the bots were trying to create. I think on one hand 2012 helped Mexican citizens to be able to identify bots and not have that be a problem. They have learned how to basically live with bots.

I think another thing that helped was that around January, the start of the year, the Cambridge Analytica story broke. What happened there was that there were allegedly some companies in Mexico that were working with Cambridge Analytica where basically they had a lot of data from people who were from low-income families, and Cambridge Analytica was allegedly going to work with that data to try to mobilize them to vote for a certain political candidate. When the Cambridge Analytica story broke, the offices of several of these companies that were working with Cambridge Analytica and also Cambridge Analytica, they had to close down in Mexico. I think that was also helpful for closing down a series of operations.

I did hear a lot of stories that different political candidates decided to be much more conservative in terms of the types of social media strategies that they used because they didn't want to be accused of doing anything related with Cambridge Analytica or those types of techniques. A lot of these companies did have to close down. I think that was also helpful in limiting the amount of misinformation that was lived in that election.

The third thing I also think that influenced it is that this was Obrador's third time.

DEVIN STEWART: Third run.

SAIPH SAVAGE: Third run. That was his third run.

I do think that the Mexican citizen became much wiser. For instance, in 2006, that was his first run, and they created a series of TV commercials where basically they were saying: "He's a danger for Mexico. Be careful. He's going to bring everything down." A lot of people did get very scared, and that was very divisive for the country.

After that, what started happening was that people also became much more critical. Now if you tell them, "This candidate is a danger for Mexico," people step back and say: "Really? Why? Tell me with information why you are saying this."

Mexicans I feel because they lived that two times, the first one was more on TV that it happened. The second time it was with the bots that came out to basically, they were defending the candidate who won. I think both of these things made the Mexican citizen much more resilient from my point of view because they were again creating a lot of spots against him. Right now, for instance, he had a very interesting activity where he had citizens vote and participate on where an airport should be constructed in Mexico.

DEVIN STEWART: Mexico City, yes?

SAIPH SAVAGE: Exactly, where the airport for Mexico City should be constructed, and people started participating. It turns out that the location that allegedly the big companies wanted, the citizens chose another one. Then they started creating a series of hashtags and commercials that, "Oh, the Mexican economy is going to go down because of that decision that he's making." Companies were trying to create fear in citizens.

What I also started seeing was that citizens were saying: "Hey, you know what? We are not falling for this." I saw on social media a lot of resistance from people. They started to mock these efforts of the companies to create fear.

For instance, they created a hashtag that was: "Oh, I'm going to protest for the airport." It was like, "I'm not going to protest for a lot of the people that have been killed by the Mexican government or [were kidnapped], I'm going to protest because I want to go travel the world to Japan." It made you think that: Oh, you know what? These companies are trying to create fear, and they're trying for us to support a cause, but they're not having us fight for causes that are more important.

DEVIN STEWART: So it's a kind of sarcasm?

SAIPH SAVAGE: Exactly. Sarcasm.

DEVIN STEWART: Okay. Interesting.

SAIPH SAVAGE: Citizens started doing a lot of sarcasm. For instance, this hashtag that they created about the protest for the airport, it was a sarcasm hashtag. That trended for several days whereas the other efforts were not able to trend. So that was interesting.

DEVIN STEWART: That's a very hopeful story what you're telling us here. If I understand you correctly, you're saying that societies can heal themselves and become vigilant against propaganda and fake news if they learn from their past mistakes or past experiences.

SAIPH SAVAGE: You're right, yes. Yes, I do believe that. You're right. It's a story of hope.

DEVIN STEWART: That's great.

SAIPH SAVAGE: Yes.

DEVIN STEWART: Maybe we could conclude with this. One question is: Do you think the same applies to the United States? That's one question.

The second question is: What are you, Saiph, doing about it to help fight fake news and help citizens?

The first one is you gave a great story about Mexico improving its society and citizens becoming more aware and vigilant and more critical thinkers. My informal amateur perspective on this is that the same thing seems to be happening in the United States just from conversations I've had with people who are in the media or watching the media and also watching the way people behave on social media. There seems to be a lot more skepticism and more critical thinking and not taking things at face value, necessarily.

What's your observation about what's happening in the United States?

SAIPH SAVAGE: I completely agree. Even in just these past elections we had a much larger number of participation, which is great, because it tells you that citizens are activating themselves.

DEVIN STEWART: Good point.

SAIPH SAVAGE: I was living in West Virginia for a long period of time, and I started noticing that citizens there were also very much as well getting involved in more political aspects. West Virginia is a state that in certain cases is not activist at all, so it was very neat to see that citizens as well were also getting involved in different issues.

It can be motivating for citizens as well to show that the stereotype that they had of them is different. I do also feel that sometimes we're thinking like, Oh, citizens are not smart, and they can't detect these big stories. It's very empowering for citizens to say: "You know what? I can detect that this is fake."

I do see a lot of people coming up and even just participating more. I think that the United States is headed in that same direction. That's pretty motivating as well.

In terms of what we're doing, within my research lab we have been building systems where we're using artificial intelligence (AI) to be able to detect biases in news stories.

DEVIN STEWART: Is that at Carnegie or West Virginia?

SAIPH SAVAGE: This particular piece was done with West Virginia University. We do also collaborate with some universities in Mexico, which is the National Autonomous University of Mexico (UNAM). With Carnegie Mellon University we also have some other projects specifically related to crowd work, which I'll also talk about.

Within the news biases, what we're doing is we're using AI to go through a series of news stories, let's say, about a protest. And then we're analyzing: "Okay. Which stories had similar images? Which stories used different images?" Then we're putting humans in the loop to help humans detect, to tell humans: "Okay, you know this story about this protest? Somehow it just presented a lot of blood. But these stories, they were showing women chanting and smiling. So this story that's showing a lot of blood might be the one that is biased. Maybe for you, a human, it might make sense to highlight that within your network."

What we're doing is we're using AI to tell humans that here there might be a bias and then letting humans decide whether or not they want to take action on that bias because, for instance, if you see that the protest that you participated in is only being portrayed by certain news media—and you might not have realized it—as a very violent protest, you may want to get on Twitter or you may want to make some calls to let people know that: "You know what? Maybe those stories are fake. That was not what happened in the protest." So we're leaving it up to the human to decide how they want to take action based on the bias.

DEVIN STEWART: Let me ask you a psychological question right there. It's very interesting. You mention two hypothetical stories, one where you went to a protest and it was peaceful and then the media portrayed it as violent. In the other one, you said there was another hypothetical story and there was a gathering, and then suddenly it was being portrayed with bloodshed.

There is a psychological concept of fabricating memories or stories where the addition of something is the thing that should be seen as suspicious. If several news agencies are portraying it one way and one happens to portray it the way the others are portraying it plus something new, then the thing that's added should be examined more carefully.

Is that true in your experience? Does that sort of psychological idea play out in your work at all?

SAIPH SAVAGE: I like this notion about looking at what is the additional thing. Here we were basically identifying when there were differences in the images, but I think that's a very interesting direction in looking at what was the new element that was added into that news story.

In this initial work that we were doing, it was just: "Okay, these images where a woman's smiling; this one was blood. Here we might have a bias, and this might be something that you want to talk about with your social network, especially"—and we would also highlight the people who were discussing, let's say, the news that was around blood so that the person could decide if maybe he wants to let these people know that: "You know what? This protest was actually something very different."

But I like that notion about having people examine what was the added thing.

DEVIN STEWART: Does it seem to ring true with your research, looking for the additional thing? I think of sort of a stereotypical crime story. Sometimes people imagine there was a weapon there, for example, and there was no weapon.

SAIPH SAVAGE: We were looking at visual bias. This is how images were presented differently. From our experience it was more that the stories were actually very different based on the images that were used, and we looked at the case of—here we were collaborating with UNAM, that university in Mexico, and we were collecting a lot of . . . The government of Mexico decided to raise the price of gasoline. Some news stories focused on people who were so upset with the government for doing that that they went to protest on the street. They said there like 100 protestors, and they would show those pictures.

Other newspapers that were more in support of the government said: "You know what? There's nobody on the street. Nobody supported that." Those were the types of biases that we were detecting. It wasn't in these cases but it might have also been because it was based on the Mexican media. Here it wasn't as much an addition but rather a different way of talking about a story.

Others were taping a protest in different locations so that they showed different numbers. But it could also be how the news media is working in Mexico, where you have groups that really clearly support the government, and you have other newspapers that are a little bit more activist. But you also have to consider that in certain cases they try to be neutral, but they are inclining more toward opposing the government, so they can also be a little bit biased as well.

DEVIN STEWART: Of course. Before we go, Saiph, do you have any final pieces of advice to our listeners on how they can become better consumers of the media and more aware and more educated?

SAIPH SAVAGE: That's a great question. I like what you said about hope. I think that right now there's a big narrative around, "Oh, we're doomed as a society because we don't know how to handle misinformation," but I think that it's important to keep in mind this notion that we have hope. Other countries have experienced it as well, and they have been able to overcome.

To be honest, I lived the election in 2006 with Obrador, and I remember that at the time I didn't think that Mexico would overcome it because the media created a big division between the people. So now it was very surprising to see those people who were once very divided against him go out and vote for him because they had just reflected that maybe they had been manipulated.

I like what you said about hope. We have hope.

DEVIN STEWART: Saiph Savage, it has been really wonderful speaking with you. Saiph is affiliated with West Virginia University and Carnegie Mellon University, one of our many Carnegie families here at Carnegie Council. Great to speak with you today.

SAIPH SAVAGE: Thank you for having me. It's an honor.