If you’ve recently discussed politics on social media, the chances are high that you argued with a robot. Every day, hundreds of thousands of political conversations take place on Twitter, Facebook, and other social networks. Most people, however, aren’t aware that a considerably high number of interactions are not with actual flesh-and-blood people, but with robots – also referred to as “bots.” They are being used to direct public debate on social media, and influence the minds of social media users – that’s to say, pretty much everyone with an Internet connection.
The 2016 U.S. presidential election showed us how social media can affect – and in some cases, decide – electoral results. Google and Facebook – the general public’s main sources of information – tend to customize what information we receive. The result is that they show us only what we want to see, and this creates a distorted and unbalanced view of how we perceive reality.
During the bloody campaign between Donald Trump and Hillary Clinton, bots were able to affect the flow of information on social media, shaping the debate and widening the divisions in American society. During the race, bots focused on making the Clinton email server scandal one of the most talked-about subjects, regardless of all the scandals surrounding Trump.
Next year, Brazil is on track to experience a similar phenomenon. This is especially the case because, according to ComScore, Brazilians spend 25 percent of their online time on Facebook, and 49 percent of us get our news from the Internet. Bots have operated heavily in our country for the past three years, especially during three particular episodes: the 2014 presidential election; the 2016 impeachment process; and the 2017 general strike.
A study published in June by Dan Arnaudo, a cybersecurity fellow at Instituto Igarapé and the University of Washington shows how bots have been present since 2014. Although Facebook is the most affected social media platform, Arnaudo focused his research on Twitter because Facebook refuses to share any sensitive data – which renders this type of research nearly impossible.
According to a political scientist interviewed by Arnaudo, social bots have been around for quite some time:
The use of bots is not something that just came about. They have been working for at least six years here in Brazil, and are increasingly common. Now, bots are becoming more sophisticated, technology is becoming more sophisticated – such as cyborgs that are a mixture of human and bot, which is more efficient than bots.
How bots try to affect discussions
Bots try to influence the popularity and public perception of certain topics and hashtags. Evidence of their presence is seen when a political hashtag abruptly starts trending despite similar subjects not following the same pattern. The first well-documented case of this occurred during the debates in the 2014 election runoff stage, between then-incumbent Dilma Rousseff, fighting for reelection, and Senator Aécio Neves.
Minutes after the start of a debate, hashtags in support of oppositionist candidate Aécio Neves would rise at an abnormally high rate. “[That’s] a strong indication that bots were being used, especially when rival hashtags supporting President Rousseff did not increase anywhere near an equivalent rate,” Arnaudo says in his study.
Companies also use bots to monitor user data and use it with the goal of pandering to different groups of people in specific ways. That helps to turn undecided voters into supporters.
The same strategy was deployed in 2015 and 2016 to influence the public mood against then-President Dilma Rousseff. Of course, as the leader of an ineffective government notable for its disastrous economic policies, Rousseff certainly didn’t do herself any favors. But the point is that bots helped to create an even more toxic environment, gathering popular support for Rousseff’s impeachment.
The 2018 election is on track to become the most divisive in Brazilian democratic history. Radical candidates, like the extreme-right-winger Jair Bolsonaro, are polling as highly competitive presidential hopefuls, while polarizing figures like former President Lula da Silva and São Paulo’s Mayor João Doria are also hoping to join the race. And they will all lean on bots for support.
Eyeing next year’s election, Cambridge Analytica, the controversial consulting company that assisted Donald Trump with his ascent to the White House, has opened an office in São Paulo. Cambridge Analytica masterfully targets voters to share content that benefits their clients – and hurt opponents. The company claims to have “up to 5,000 data points on over 230 million American voters.”
But the company will face local competition. In Ouro Preto, Minas Gerais, a group of 35 developers from a company called Stilingue has spent the last four years creating a software project called War Room. It monitors social media discussions and user behavior, with the goal of helping campaigns pander to different demographics. Stilingue’s CEO told the BBC that two presidential hopefuls have already subscribed to their services.
How to fight against the machines
Bots are already a part of our contemporary political landscape. While their use has been more widespread in right-wing factions, make no mistake: they are indeed weapons used by all major parties. And Brazil’s electoral justice system has thus far failed to keep up with how campaigns are using online propaganda to their advantage. And it is likely going to get harder, as Congress has approved a bill allowing candidates to sponsor their content on social media.
These guerilla tactics are often employed by companies with no formal links to campaigns. While the law forbids candidates from campaigning outside of a time period established by the Superior Electoral Court, no laws exist to prevent private individuals from supporting a candidate through their social media profiles. And that’s how campaigns are able to cross the line.
Dan Arnaudo told The Brazilian Report that users must learn how media and information spreading work. “People must understand what they’re reading, what the source is, why – and if – you should trust it. Users must be more careful with what they share on social media to avoid spreading unverified information.”
Arnaudo also holds tech companies accountable for how their networks are being used to manipulate voters and spread fake news. “Corporations, especially Facebook and Twitter, must find and block [bot] accounts instead of supporting these kinds of strategies – like they did until 2016.” He also adds that political parties should denounce the use of bots, and pledge to avoid engaging in such tactics.
Unfortunately, these potential solutions seem like a long shot in our current climate, especially because polarization generates user engagement – and that’s a magical phrase for social media managers these days. Facebook’s auction-based system rewards ads that generate more likes, comments, and shares; since polarizing posts are more likely to attract vehement supporters and detractors, they become effective engagement strategies. Provoking your enthusiasm or outrage is the business model right now.
As we move into the 2018 election cycle, just keep in mind that now-iconic phrase: “If you’re not paying for it, you become the product.”