The details of the plans went like this: After the election, the government will take full control of the internet. It will establish a stringent schedule of times citizens will be allowed online and will restrict access to websites it considers harmful to society. According to the new rules, the feds will gain the power to remove any media they consider false, biased, or defamatory. In order to eliminate obesity, junk food will be prohibited, and each person will be given a monthly cap on the amount they spend on meals. No private sector employee is to make more than $800 per month. No individual will be allowed to own more than one car.
This isn’t the premise for a dystopian novel. These are the plans for how Mexico’s President-elect Andrés Manuel López Obrador (better known as AMLO) wants to start his six-year term—at least according to a document spread widely on WhatsApp, Twitter, and Facebook after the leftist, populist candidate recently won Mexico’s election. The bulletin was, of course, fake. And it provides just one example of the many false news stories debunked by Verificado (“Verified”) 2018, an independent, collaborative fact-checking initiative that worked full time in the lead-up to the country’s July 1 elections.
Operating alongside established networks of corruption and violence, fake news proved a formidable enemy to a free, fair, and informed vote in Mexico. It also made the country’s election the most recent battlefield in a global fight against digital disinformation and misinformation. While the purveyors of false information in Mexico often took advantage of discord specific to that country’s electorate, Mexico’s election nevertheless provides a revealing lens into the way this kind of weaponized media has evolved and adapted. As companies like Twitter, Google, and Facebook seek to address the spread of false information on their platforms in other large democracies like India and the United States, Mexico’s experience seems telling.
As in the 2016 U.S. presidential election, the runup to the 2018 election held by our southern neighbors also saw complex online networks of false information disseminated by bots, trolls, political parties, and individual citizens caught up in a climate of distrust and frustration. Despite worries that Russia might execute large-scale meddling in Mexico as it has in major elections in other countries, the Atlantic Council’s Digital Forensic Research Lab was “unable to verify claims of large-scale bot activity emanating from Russia,” though it did find a commercially run botnet, apparently from Russia, that boosted political messages in Mexico. More broadly, it appears the largest sources of false information came from inside, not outside the country. Most false stories revolved around presidential front-runner AMLO. Still, examining the fake stories that spread during and immediately after the election, it’s clear no candidate was safe from falsehood.
We know a lot about the form and spread of this false information thanks to Verificado 2018, which was started by a team of journalists from Pop-Up Newsroom, AJ+, and Animal Político in March with the mission to detect and debunk popular fake news, memes, and images related to the election as well as to fact-check statements from major political candidates during the campaign. The project quickly expanded to include more than 80 allies in local and national media outlets (plus some universities and advocacy groups), all of which coordinated with Verificado to research and publish particular verifications at the same time. By the time the group stopped generating content just after the election, it had produced more than 400 notes and about 50 videos examining false claims, documenting suspect sites, and investigating how certain fake news went viral, according to its farewell blog post.
Verificado’s archives offer a rich sampling of the kinds of suspect stories that took off on Mexican social media. On Facebook, for example, a post falsely claimed that AMLO’s wife was the granddaughter of a Nazi. A meme that circulated on various platforms, including the hugely popular WhatsApp, urged anti-AMLO voters to check boxes for both his opponents in order to ensure the leftist would lose—an action that, in reality, would nullify their vote. Another set of popular posts repurposed photos of porn actors to depict a hitman and a multimillionaire who were supposedly conspiring against AMLO, one of many posts seeking to spread the idea that a “mafia of power” was working against the candidate. One doctored video that made the rounds appeared to portray the pope condemning AMLO. Others claimed that the Venezuelan or Russian governments had rallied behind the candidate. At one point, the Russian rumors became so pervasive that AMLO released a humorous video of himself standing on a dock, where he jokes that he’s waiting for a Russian submarine to bring him gold from Moscow and that he’s now “Andrés Manuelovich.”
At times, disinformation also helped to paint positive images of the candidates. For example, one post purported to show a massive stadium of supporters who had come to watch AMLO close his campaign in the state of Guanajuato. But it turned out the photo was actually lifted from a 2013 Green Day concert in England. Similarly, a photo captioned, “A tsunami named Andres Manuel Lopez Obrador arrived in Quintana Roo,” actually shows the crowd from a 2006 Rolling Stones concert in Rio de Janeiro. AMLO, though he may have been popular, has a hard time competing with rock star–level audiences.
Fake or manipulated polls played a similar role. A week prior to the election, leaders from the National Action Party shared a poll showing their candidate, Ricardo Anaya, seemingly closing the lead on AMLO. (While Anaya ended up coming in second, there was a 30-point gap between the two candidates.) The poll looked as though it was conducted by El Universal, one of Mexico’s top newspapers. True, the poll did appear in El Universal—but as a paid advertisement placed by the company “Pauta Encuestas” with no attributed methodology. Additional fake polls—including one from April that gave AMLO an astounding 78 percent of the vote and one from March that placed José Antonio Meade, the eventual third-place candidate, at 42 percent—were falsely attributed to the New York Times and shared by thousands.
These false stories represent just a few of the narratives the Verificado 2018 team debunked. Some of the posts came from websites or social media accounts Verificado designated as principal generators of false information—many of which still publish and share content with hundreds of thousands of followers. These kinds of stories spread on some of the platforms that typically see circulate information in the U.S., like Facebook, Twitter, and YouTube. But some also did stunningly well on a different frontier: the encrypted messaging service WhatsApp. According to the Reuters Institute, 35 percent of Mexicans use WhatsApp to get news. WhatsApp’s popularity can be particularly problematic when it comes to the spread of false information, because the platform’s danger lies in how personal it is—when users spread misinformation, they’re usually doing so by sending it directly to those they trust (though misinformation also spreads contagiously in WhatsApp’s ubiquitously used group chats). This not only means the bogus information is more likely to stick, it also makes it harder to track and counter than when the same stories get shared on social media platforms.
To address this type of fake information, Verificado set up a WhatsApp hotline where users could send questionable messages. A Verificado staffer would then respond directly to most messages as well as post statuses about stories they had debunked throughout the day. According to Verificado, more than 10,000 users subscribed to this hotline. On July 10, in order to crack down on deadlier forms of disinformation being spread on the platform, WhatsApp also announced that it would begin labeling forwarded messages—which it says will help users determine whether the messages they receive were written by their sender or are part of a larger forwarding chain—and restrict message forwarding in some regions.
To be sure, Verificado’s outreach effort likely reached just a fraction of Mexico’s 89 million eligible voters. But as Alfonso Méndez Forssell, a media researcher who also worked on a digital communications project supporting AMLO, told me, Verificado’s services did play an important role in opening the public’s eyes to the scope of disinformation. They also created demand for “a climate of rigor, of accountability,” he said. He thinks news organizations and digital platforms should take some responsibility for teaching voters what disinformation looks like and how it spreads. But, he said, he also thinks that schools should make media-literacy education mandatory in order to ensure the next generation is equipped to make informed decisions.
In Mexico, that matters beyond the political sphere. While Verificado 2018 focused on disinformation in the election, its roots can actually be traced back to the deadly Sept. 19 earthquake that rocked central Mexico and killed about 370 people. The aftermath left survivors lost in a climate dominated by confusion and fear. Amid this, false information began circulating in social media posts and WhatsApp messages, including rumors that authorities were predicting a larger earthquake was coming soon and claims about which buildings had collapsed or were in danger of collapsing. The effect, as one reporter for the newspaper El Economista put it, was a “psychosis within an already overwhelmed society.” To combat some of this false information after the earthquake, a group of journalists, activists, and programmers founded Verificado19S, which used a team of fact-checkers, and Google Maps, to organize information about where donations were needed, which areas were actually hardest hit, and how volunteers could help.
An election is different than a natural disaster, but false information can be destructive in both contexts. On its website, Verificado 2018 notes that like an earthquake, false information “confuses, paralyzes, frightens. It seeks, in that sense, to undermine citizens’ freedom to decide who to vote for.” Mexican citizens know it: According to the 2018 Edelman Trust Barometer, about 80 percent of Mexicans are worried about “false information or fake news being used as a weapon.”
Of course, none of this is entirely new. As Jorge Buendía, a global fellow at the Wilson Center’s Mexico Institute, points out, “The use of fake polls has become the rule rather than the exception for electoral processes in Mexico.” Candidates have also used bots and paid trolls to influence Mexican elections since at least 2012. But today, what’s different is the scale and speed at which such disinformation can reach the electorate. As the New York Times reported, there are now about 71.3 million internet users in Mexico, whereas in 2012 there were only 40.9 million. What’s more, this near doubling of internet penetration didn’t come with corresponding efforts to teach digital media literacy.
This rapid change—on top of increasing political polarization and a general distrust of government among the Mexican electorate—allowed false information to be planted in what Buendía called “fertile ground” for fake stories. According to a poll of 1,003 Mexican adults he conducted in November, 53 percent of respondents believed there was a secret plan to prevent AMLO from becoming president. About 29 percent said they believed the Mexican government knew about the country’s deadly Sept. 19 earthquake in advance but decided to keep that information from citizens. With this kind of discord to play with, actors wishing to manipulate parts of the population for political or economic gain will surely continue to seed the same kinds of disinformation.
But, as Mexico proves, there are ways to fight back. Though Verificado 2018 has closed, some partner organizations will continue to carry out similar fact-checking initiatives, recognizing that these false stories won’t just pop up during presidential campaigns. AJ+ Español, for example, recently launched AJ+ Verifica, which will continue to verify suspicious content through its social platforms. Like ProPublica’s Electionland project in the U.S. (which recently relaunched for the 2018 midterms) and CrossCheck’s initiative in France and the U.K., such collaborative journalism projects may provide one model for keeping some digital disinformation in check.
Despite all the challenges and harm wrought by false information, it does present a small opportunity for media organizations to regain some public trust, explains Juliana Fregoso, a journalist and digital news initiative fellow at the Reuters Institute.
“My impression is that in Mexico, the media stopped writing for the reader a long time ago,” Fregoso wrote in an email. “And this is a good opportunity for the media to start again with journalism that people are interested in reading, not journalism that interests politicians or powerful leaders.”