Astroturfing with Patrícia Campos Mello
Season 7 Episode 3 [download this episode (MP3, 586 MB)]
In this episode, Elizabeth interviews Brazilian journalist and columnist Patrícia Campos Mello, who first exposed the illegal use of WhatsApp mass messaging in Brazil's 2018 presidential elections. Together, they examine how politically motivated groups orchestrate fake grassroots movements—known as astroturfing—to manipulate public opinion and influence media coverage. Patrícia shares Brazilian examples to highlight the evolution of these tactics, from mass messaging operations to highly coordinated content promotion strategies. They also delve into the ethical and regulatory challenges that astroturfing presents for democracy.
Additional resources:
Elizabeth defines astroturfing based on Digital astroturfing in politics: Definition, typology, and countermeasures by Kovic et al. (2023) and Online astroturfing: A problem beyond disinformation by Chan (2024).
For additional context on the 2018 Brazilian Presidential Elections, consult: Vox's explainer on Brazil's 2018 Presidential Elections
Patrícia discusses the tools and strategies used for mass messaging campaigns. For a more detailed explanation, consult: How WhatsApp is being abused in Brazil's elections
Patricia's original report exposing the illegal use of WhatsApp mass messaging practices is available in Portuguese with an English summary.
Follow Patrícia on Instagram @patacamposmello and X @camposmello.
Contributors:
Host: Elizabeth Dubois
Research Lead: Louise Stahl
Transcript: Louise Stahl
Audio Editing: Ayman Naciri
Translation: Els Thant & Helena Legault
Publishing and Promotion: Helena Legault
Producer: Michelle Hennessey
Episode Transcript: Astroturfing with Patrícia Campos Mello
Read the transcript below or download a copy in the language of your choice:
Elizabeth: [00:00:05] Welcome to Wonks and War Rooms, where political communication theory meets on the ground strategy. I'm your host, Elizabeth Dubois. I'm an Associate Professor at the University of Ottawa and University Research chair in Politics, Communication and Technology. My pronouns are she/her. Today we're talking about astroturfing. Patrícia, can you introduce yourself, please?
Patrícia: [00:00:24] Sure. My name is Patrícia Campos Mello. I am a journalist with Folha de São Paulo, that's the leading newspaper in Brazil, and I have been covering disinformation and influence operations more or less since 2014. And yeah, I guess this is it.
Elizabeth: [00:00:41] That's fantastic! I am so excited to be talking to you today. Your experience covering disinfo [disinformation] and these influence campaigns is not rivalled by many; you've been working at this for a decade at this point. So I'm very excited for your perspective. Today we're talking about the idea of astroturfing specifically, which is, kind of, in that larger umbrella of disinfo and influence campaigns. I want to start off with a bit of an academic definition, and we'll see whether or not that makes sense to you and how you understand the term. So astroturfing is a tactic that's used by politically motivated groups or players to mimic grassroots movements for their own gain. So the idea is: make it seem like there's this bottom-up swell of support from the public for a particular politician or idea or policy, when in fact, it's actually a, usually carefully orchestrated, quite centralized campaign effort meant to mimic that grassroots approach. So we get the name astroturfing from the idea of astroturf, which is, you know, fake grass.
Elizabeth: [00:01:52] In an online context, that astroturfing can be particularly deceptive [consult: Online astroturfing: A problem beyond disinformation]. And that's because it can happen at this scale, not really imaginable when we're thinking about an offline context and there's this ability for a degree of anonymity, it's usually pretty low cost, making all of it pretty hard to trace and difficult to understand who's the originator of these messages, even if we determine that they might be deceptive messages. And so when we think about astroturfing, it could be potentially dangerous because it can really distort public's perceptions of what is actually popular or credible. And we know that those things go into people's voting decisions, for example. We'll untangle this a lot more. But before we do, I wanted to check in. Is that description and definition making sense to you? Would you change anything? Any questions?
Patrícia: [00:02:48] Absolutely. I totally agree. The power of astroturfing to mould the conversation, to make it seem like everyone is talking about one particular candidate or one particular issue, is just overwhelming. We see this in Brazil, always. It happens a lot because they have been improving the method.
Patrícia: [00:03:10] I've been to India several times too, covering this. So it's also another country where they are really experts in doing this [Learn more about astroturfing in India]. So yeah, I think the definition is perfect.
Elizabeth: [00:03:20] Great. Let's talk about how that has evolved. So, we have the early versions of it that are kind of continually being improved upon as people experiment and as people try and evade detection. What are some of the early versions of it that you can think of?
Patrícia: [00:03:36] I can talk about what I've been covering now and what I covered back in 2018. In 2018, we had a very contested election. It was the extreme right-wing leader, Jair Bolsonaro, against Fernando Haddad, the Workers' Party candidate [For additional context, consult Vox's explainer on Brazil's 2018 Presidential Elections]. And what happened was, Brazil is a country that runs on WhatsApp [For additional context consult: Statistics on WhatsApp use in Brazil]. Everybody's on WhatsApp. Supreme Court judges are on WhatsApp. Every single person. It's the internet, right? The other day, I interviewed the CEO of WhatsApp and he said that, even though it's not the largest market for WhatsApp — I think the largest is India and then Indonesia —, it's the largest market in terms of sending audio messages and videos and forming groups [Consult: Patricia's interview with the CEO of WhatsApp].
Patrícia: [00:04:19] So, in 2018, what happens is that Bolsonaro, he was a pioneer in using social media for campaigning. And this is the other thing: sometimes the astroturfing is not done by the campaign itself or the candidate itself. It's always an arm's length operation. People associated or supporters or businessmen. [For more details consult Patricia's original report exposing the illegal use of WhatsApp mass messaging practices] [...] So at that time, what we had was like an assembly line of disinformation working to boost Bolsonaro's candidacy and disseminate false information about his opponent. So it was literally an assembly line in that you had marketing agencies that were selling services.
Patrícia: [00:05:05] At that time [it] was very primitive because you had either these adapted machines in which you inserted SIM cards. Or afterwards, it got a little bit better and you didn't need that, you had only a dashboard. But basically what you did was you bought in the parallel market lists of phone numbers of voters. And this was divided by profiles of people you were intending to reach. So, you could buy urban people, middle-aged women, or conservative, etc. And I have to say, that's what they said the data was. It wasn't always that efficient. But that's what they were selling. So the candidates or their supporters, they bought all these lists - and this is illegal in Brazil, they couldn't buy from the marketing agencies as well - to send millions and millions of messages to these people and to public WhatsApp groups [For more information on how data scraping and mass messaging work consult: How WhatsApp is being abused in Brazil's elections]
.
Patrícia: [00:06:06] So, what happens is that we had this one particular piece of disinformation that was called the Mamadeira de Piroca, which was the "penis-shaped baby bottle." So the story was that the Workers Party candidate was going to introduce in public schools the penis-shaped baby bottle, because this was part of his plan to teach gender identity and to incentivize kids to gender transition. So, this was all made up. It got viral in a way. All the groups had that. Most of the disinformation is the classic "elections were fraudulent." They show you a voting machine somewhere that is not in Brazil, saying, "This machine is in Brazil and they're not registering the right votes, etc." [For more context, Consult: Brazil election: Do voting machines lead to fraud?]. But some of them was a cultural wars thing [Learn more about the concept and How the ‘Culture War’ Could Break Democracy]. And the fact that they could hire marketing agencies to send out millions and millions of messages, and these services were bought by Bolsonaro supporters. So, this was always not directly linked to the campaign. Anyway, so this was back then.
Elizabeth: [00:07:12] Yeah, it's a fascinating example and really points to one of the three dimensions that Kovic and colleagues [Consult: Digital astroturfing in politics: Definition, typology, and countermeasures] talk about, which is: who is the main actor? We think about the actor, the target, and then the goal. And so, the idea of the main actor is not always who you would expect the actor to be. And then, how do you trace who that is? And why is it possible to have these actors who are arm's length or multiple steps out who are able to send these political messages that are very clearly partizan in nature?
Patrícia: [00:07:44] And also, this is a way of bypassing electoral regulations as well. Because, in theory you need to declare, to have transparency of all your electoral spending to the electoral authorities. Astroturfing is also a way of bypassing all electoral regulations. Because it's not you doing this. Someone is doing on your behalf and actually paying for people to pretend that they are either supporting you or to disseminate lies about your opponent. So, at that time, it was not illegal to send mass messaging. That's how we call it: WhatsApp mass messaging. And then, two years later it became illegal [Consult: Brazilian General Data Protection Law]. It's a very complex discussion. People ask, so did it make a difference in the election? I don't know, we'll never know. If it just crystallized people who already had intention of voting for someone or, we don't really know. But the fact is that they were very successful in moulding the news cycle and sort of dictating what was going to be the mood and the topics discussed. You just think about it: "penis shaped baby bottle." What the hell? And then it's just like everybody's talking about this.
Elizabeth: [00:08:55] And it really just sets the agenda. This has become such a cultural phenomena at this point, because so many people received the message that you can't not talk about it.
Patrícia: [00:09:06] Exactly.
Elizabeth: [00:09:06] All right. So, that was a 2018 example. More recently, how has it changed? What does it look like now?
Patrícia: [00:09:12] This year, there was something that was really fascinating. These municipal elections. We just had the first round of municipal elections on Sunday, October 6th. And one of the candidates [for mayor of São Paulo] was this really famous internet influencer [Pablo Marçal]. Like the guy who is a coach, and he gives this, self-help courses, and he has millions and millions of followers on TikTok, on Instagram. So what this guy did, before he was a candidate, he had contests where people were told to get his videos and make edits of the videos, and the ones that got the most views would generate awards in cash for the people. So basically, he had thousands of people who were getting his videos and photos, making these edits the best they could so that it would go viral and they would get the award, the cash award. And that continued into the election. He had an army of people just flooding social media with his videos, with his edits, and there was no legislation for that. The electoral authorities had to adapt to that, saying: this is electoral advertising. This has to be counted somehow. And you can't just pay people to disseminate videos of you. But this was another type of astroturfing, because if you are someone who doesn't know that [is happening], you are just looking at social media and the guy is everywhere.
Patrícia: [00:10:47] We had a virtual [statistical] tie among the three top candidates. I mean, it was very, very close. This guy who came from this party, that's insignificant, he got in third, but he almost got to the second round; and he had no time in electoral advertising on TV [Editor's note: in municipal elections in Brazil, TV time is partially split among candidates proportionally based on their party or coalition's representation in the Chamber of Deputies]. It was just the video stuff. His people just spreading videos.
Elizabeth: [00:11:08] Yeah, that is such a fascinating example. And it also starts to pick at some of the issues we have with determining what actually counts as astroturfing. Because, in this case, they were actual humans, actual potential voters who were making the videos and sharing them. So it's not fake creators of content, but it's not necessarily them supporting that idea. It's them wanting to win money.
Patrícia: [00:11:38] Yeah, because they were actual humans. But they were not necessarily his supporters. They were there for the money. They wanted the cash award, so it could be anyone.
Elizabeth: [00:11:49] So, it's like, the money, the incentive, becomes the problem. Because on one hand you could frame it as: well, that's advertising spending that isn't being calculated and needs to be calculated. And also — I don't know if in Brazil there are requirements about statements of endorsement or approved by the financial offices for such and such a candidate — but in Canada, that would be a requirement and would be violated in this case.
Elizabeth: [00:12:13] Or, much more nefariously, it's like trying to pay people off to vote your way, which is very illegal in basically all democracies.
Patrícia: [00:12:22] Part of the videos were promising voters or participants gifts, hats, t-shirts, small stuff. But this is illegal here. To promise any kind of gift in exchange.
Patrícia: [00:12:35] And also, as for internet advertising, it can only be done and paid by the candidate, him or herself and the party. So the fact that it was his companies paying these people to spread the videos, it's a web of new things that are coming up. And then what you said, is this astroturfing or not? I mean, these people, they're real people. But are they doing this because they really support him? No, they're doing this because they want the cash reward. And then if you're a regular voter, you are on the internet. You're on Instagram,TikTok, it's like, "Oh my God, everybody loves this guy. There are so many videos."
Elizabeth: [00:13:14] Yeah, it looks like it's this grassroots campaign. Which is exactly what astroturfing is trying to do. So you mentioned, if you're in the public, you're on social media, you're seeing, "wow, all of these videos." There clearly the target for that kind of a campaign would be the general public. Do you see astroturfing used for other target audiences ever?
Patrícia: [00:13:37] You mean in commercial?
Elizabeth: [00:13:38] Well, I'm thinking like there's the general public. I imagine maybe astroturfing could also be used to try and convince particular niche groups within the public or demographics, maybe. Or maybe it's astroturfing to try and influence what news is going to cover. So, your target is actually news media rather than the general public because you want to use news media to get at the wider public.
Patrícia: [00:14:03] I think in other countries, they used Facebook ads for that. But at that time, you were able to segment more the ads. And now it's more restricted. The other thing that they were doing at one point was, you could buy those softwares that create phone numbers, users from Instagram and Facebook. Let's say I want to scrape all the users from specific groups that are focused on anti-abortion groups, let's say, and then you've got all the phone numbers and could micro-target these people. I tested, just for the purpose of the story, the software. On Instagram, everybody who posted [a particular] hashtag [...] you could target these specific people, but then it's not astroturfing, it's just micro-targeting.
Elizabeth: [00:14:52] Yeah. Probably also, the ones we're more likely to hear about are the ones that are those bigger, broader campaigns that are sort of blunt instrument, just trying to get at many people. Because those are the ones we can notice.
Patrícia: [00:15:07] Otherwise, if you're not being targeted, you won't know. And that's the whole point. So they're just improving and improving the whole thing. I mean, this thing about the contest, it's actually really ingenious.
Elizabeth: [00:15:17] Is kind of brilliant, right? And it is a way of mobilizing a big group of people to send a message.
Patrícia: [00:15:23] They had the groups for the contests on discord. So, discord they established. Okay. So, "How many views did you get and who's going to get (...)." All the organization. And then, they went on and posted on TikTok and Instagram. So it's interesting. [laughs]
Elizabeth: [00:15:40] Yeah. And you can imagine a version of that that uses the exact same strategy, minus the "We'll pay you for this." And it's more around like, "Hey, you're a supporter," you show your support for the candidate, and you orchestrate that same sort of competition, you get to be part of this discord where you feel like you're part of a community. I could imagine that strategy being used and potentially even lauded as a great way of engaging people in their democratic system.
Patrícia: [00:16:11] Yes. And to what extent that is legal or not, if there's not money involved? Is that a reasonable way to motivate your supporters? I don't know. I don't have an answer for that.
Elizabeth: [00:16:23] This is something I've been thinking about a lot. In Canadian election law, at least, there's the tracking the money. If money has been spent, you have to report it.
Elizabeth: [00:16:35] Maybe even disclosure about when money is spent for advertising. There are pretty clear bounds around that. Now, there are people who find loopholes and try to evade those laws, but it's at least something we've got a bit more of a tangible grasp on. There are rules that limit the extent to which you can coordinate there are those limits, that is something that's already ingrained in Canadian election law. And in a lot of democratic nations, their election laws. But it's so much harder to decide where normal campaigning ends and where coordination that's not good for democracy begins. It's just a gray zone that I don't think anyone's really figured out.
Patrícia: [00:17:15] And also as rules for social media platforms or their internal rules, the (...)
Elizabeth: [00:17:24] The community guidelines?
Patrícia: [00:17:26] Guidelines, as they get stricter, people find ways. Here in Brazil, legislation was implemented at the beginning of the year that demands way more transparency from political advertising on Facebook, on Google. Facebook already complied with this, with the political ads ad library, right? Google's transparency report here in Brazil, it's very deficient. Compared to the one in the European Union, there's very little they are disclosing. So they just decided, "Okay, we're no longer going to sell political ads in Brazil because it's not worth it having to comply with all the transparency requirements." [Consult: Google's decision to stop political ads] But then, at the same time, you have people using discord to organize advertising campaign that is nontraditional, but it's still (...) Where's the transparency? There's none.
Elizabeth: [00:18:17] I've personally been really interested in how social media influencers are being integrated into campaigns, because I think there's similar things. Many of them have their own discord that they use in that way as well. But there are these networks of influencers who all start to share particular political ideas and messages, and maybe they've been paid off platform. Maybe they haven't. We don't understand really what's going on. It's all a little bit murky territory, but they have a really powerful connection with diverse audiences. And so that's a way of getting around the fact that political advertising is not officially, formally allowed on whichever platform.
Patrícia: [00:18:57] Yeah. The influencer component is just (...) How can you prove if an influencer is just legitimately supporting a candidate, or if he or she is getting paid? If she's being paid, should she disclose that to the followers? It's not, "I really love this candidate."
Elizabeth: [00:19:14] Yeah, exactly. Should they disclose it to their followers? Do they have to disclose it based on the platform that they are putting the message out on? Do we have some right to know when they are doing that in the more private spaces that have been created, like invite-only discords or the special tiers of Patreon or what not? Yeah, there's a whole world in this social media ecosystem that I think we aren't paying attention to because they are the slightly private spaces. WhatsApp is a really good example of that, that we have at least been able to start to understand a bit better. But there's a bunch of other venues that this is all happening on.
Patrícia: [00:19:55] We have now some tools to follow or monitor the WhatsApp public groups [Consult: WhatsApp Monitor: A Fact-Checking System for WhatsApp]. Some of the groups, they just share links and ask people to join [Editor's note: WhatsApp groups can be made so that anyone with its link can join. In Brazil, these links are often shared on other groups, social media or directories]. Some of them are political groups and since they're open, we can join the groups and see what's going on. But what I was thinking is the idea of being everywhere at the same time, which was what happened with the disinformation spread by WhatsApp. And now with the video edits, it's just like they managed to be everywhere.
Patrícia: [00:20:30] And that's a fake perception because they don't have supporters everywhere who are declaring their passion for the candidate. That's why I think it is astroturf.
Elizabeth: [00:20:41] Yeah. Absolutely. Okay. So at this point we've talked about some of the different actors involved, the different kind of targets. Let's talk a little bit about that third dimension: the goal. What in your view is typically the goal of these kinds of astroturfing efforts?
Patrícia: [00:20:56] I think it's to manipulate perceptions. The voters. Media. The news cycle. It's quite successful.
Elizabeth: [00:21:07] Yeah, it gets lots of attention. People need to report on it, once it's gotten big enough because it is legitimately a part of that electoral campaign or even outside of an election, we know that these kinds of things happen. It's interesting to me when I think about the other research on disinformation strategies, and there's some disinformation tactics where it seems the goal is to make you not really trust your information environment, to not know what's going on at all [Consult: Online astroturfing: A problem beyond disinformation]. To kind of create a bit of chaos. And then there are other strategies, and I would venture to say most of the time when astroturfing is used, which are more about creating either support or opposition for a thing. But maybe I'm wrong. Maybe astroturfing also creates that sense of "I don't even know" in the environment.
Patrícia: [00:21:58] I agree with you. Sometimes it's not even like being a pro or against a specific candidate. It's just flooding social media with random stuff. Or things that people get confused, they don't know anymore what to believe in. [...] Here, two days before the election, the same guy - the influencer whose name is Pablo Marçal, by the way - he published in his social media, and - of course - it was disseminated by the followers who participate, a fake medical statement saying that one of his opponents was using cocaine [Learn more on the sharing of the false medical report]. It was manufactured. And he put it there for, I don't know, two hours? And then the platform [..] took it down. But, at that point, it's everywhere.
Elizabeth: [00:22:48] Yeah. And it's the kind of thing that's shocking. It's going to get people talking and be, "Oh, right." And when you're right before Election Day, there's not a lot of time to fact-check and get the correct message out, and to clarify things.
Patrícia: [00:23:03] The media did fact-check. [...] The doctor was dead. The signature was fake. The clinic was this shady place. But then again, I don't know if everybody who got the first one [the disinformation] that he [Pablo Marçal] has this huge chain of, this army of online disseminators, got the actual news [the correction], the same amount of people.
Elizabeth: [00:23:28] Absolutely. And we do know from fact-checking academic studies at this point that the fact-checked rarely reaches as far as the initial deceptive or disinfo piece of content [Consult this summary on current approaches to fighting misinformation and new alternatives]. So it really does create a problem there, particularly when we're thinking about something that is so time-bounded like an election.
Patrícia: [00:23:51] Just one thing that might be an optimistic note. In that case, even though he did this fake medical statement, it was so obvious that it was fake, and there were so many evidences that it was fake, that political scientists actually think that it might have backfired [Consult a preliminary analysis of the backfire effect. Text in Portuguese]. And the other candidates managed to make the accusations that he forged a medical statement go viral. And in that way, he lost because of that. So, maybe we can counter that with the truth somehow. I don't know.
Elizabeth: [00:24:28] That speaks to why it's important to do fact-checking, even if not all fact-checks reach as far as one would hope. Sometimes they do, and it can make a difference. So there's fact-checking. You mentioned Brazil updated some of their laws over the last few years. Do you think there are ways to effectively regulate or prevent these things? Are the laws that have been updated sufficient?
Patrícia: [00:24:57] They're always going to be one step ahead of the legislation, but you have to have legislation. In that case, it was something that no one thought about. But then you just create judicial precedent for that. It's always that the legislation is going to be behind the technology. At least some disincentive so that maybe they are not so open about doing it. I don't know.
Elizabeth: [00:25:24] When we think about how political communication tactics evolve over time, there's always a bit of that push back and forth. Campaigns are always trying to get the edge. Their supporters are always trying to get the edge over the others. We always see, "How far can we push the bounds?" "Okay, now we got to pull back." So, if law or journalistic coverage or people's media and digital literacy aren't advancing with it, then it could run a lot farther down. By having at least a bit of back and forth with these countermeasures, it keeps it a bit more within control, I suppose, even though it's not 100%.
Patrícia: [00:26:02] One of the things that changed here, that maybe is making the tech platforms a bit faster to act. Because we just had a coup attempt in 2023. And just like the attack in the capital in the US in 2021, it was mostly organized by people on social media, incited by political leaders that actually also use social media. So the same thing happened in Brazil [Learn more on the Brazilian Capitol attack]. There's even a Facebook Oversight Board decision on that. Showing that the moderation policies failed, in the case of Brazil. There were things that were clear glorification of violence or inciting violence, and nothing was done. And this time, the government is implementing some regulations, that make platforms a little bit responsible. It's not a change in our section 230 yet [Editor’s Note: Patrícia is talking about section 230 of the US’ 1996 Communications Decency Act]. But [there are] some changes there. I think maybe in other circumstances, Facebook, Meta would not have been so fast in removing the fake medical statement. But now I do think that it [the new legislation] works a little bit in making them faster in acting when there's such a campaign going on.
Elizabeth: [00:27:15] We know that speed matters a lot for catching these kinds of things and fixing them to the extent that we can. There is so much more that we could talk about, but we are getting real close to time. So I want to finish off with the little pop quiz I end each episode with, and that is, can you give me the short answer exam response for how do you define astroturfing?
Patrícia: [00:27:42] Mm. I think astroturfing is the artificial coordinating social media movements to manufacture grassroots support. [Editor's note: Astroturfing is not exclusive to social or digital media, but as explained by Elizabeth in the beginning, this medium brings new challenges of scale and anonymity.]
Elizabeth: [00:27:57] Yeah, I think that's pretty spot on. It's this manufacturing of grassroots support, using social media to try and make it look like there's all this grassroots support when actually there isn't. Wonderful. Thank you so much. I really appreciate you taking the time. This was a great conversation.
Patrícia: [00:28:15] Thank you. Thank you for having me.
Elizabeth: [00:28:17] All right. That was our episode. Looking at astroturfing. I hope you enjoyed it. As always, you can find links to the things we discussed and additional resources in the show notes or in the annotated transcripts available in English and French over at polcommtech.ca. I also want to acknowledge that I am recording from the traditional and unceded territory of the Algonquin people, and I want to pay respect to the Algonquin people, acknowledging their long standing relationship with this unceded territory.
