It is nothing new that malicious automated profiles have invaded news feeds and social networks, delivering a pig in a poke or, worse, lots of lies in place of the truth.
In the United States, political turmoil was the ideal stage, with an unpredictable end and scary trade-offs. In Europe, the attention over this issue has intensified in the post-American election era. And yes, the problem has also reached Brazil, which has elections scheduled for 2018.
A study published at the end of August (in Portuguese with illustrations) by the Public Policy Analysis Direction of the Getúlio Vargas Foundation (FGV/DAAP) analysed the illegitimate interference in the public debate on the Web as well as the possible risks to democracy and the Brazilian electoral process in the coming year.
The massification of false posts has become a potential tool for the manipulation of debates in social networks, especially in moments of political relevance.
According to the study, for example, in the national general strike of April 2017, more than 20% of the Twitter interactions among users in favour of the strike were triggered by this type of account — one robot programmed to interfere in the debate for every five real users. During the presidential elections of 2014, robots also managed to generate more than 10% of the debate.
But the FGV study goes further and shows that robots are as polarised as the discussions in Brazil. They navigate on the two ideological extremes in the country. They are neither left- nor right-wing. In the last presidential election in 2014, 9.7% of the tweets in support of Dilma Rousseff came from automated accounts. For Aécio Neves, the number was 19.7%. In the impeachment, 21.4% of the support for Dilma came from robots.
Robots also take advantage of the fact that, generally, people have no criteria when following a profile on Twitter or accepting a new “friend” on Facebook.
A large-scale social bot infiltration on Facebook showed that for than 20% of legitimate users accept friendship requests indiscriminately, according to the article “The Rise of Social Bots,” published in the Association for Computing Machinery magazine. “Sophisticated bots can generate people that appear as credible followers, and thus are more difficult for both people and filtering algorithms to detect,” the study reported.
In a country like Brazil, where almost half (49%) of the population uses the Internet as the primary source of information (according to PBM data for 2016), people must constantly remain alert. It is more necessary than ever to understand, filter, and know how to report the use and dissemination of manipulated news, including by publishers. Newsrooms also need specific training to learn how to identify and map mechanisms that inflame counterfeit news.
The best way to combat the army of malicious robots is accessing qualified information, including information about fake news practices. It is important to be clear about the performance of these robots that not only disseminate false news with very harmful effects to the democratic society, but also actively seek to prevent users from informing themselves properly.
This is where the well-known platforms and search engines need to invest, ensuring more effective distribution of qualified content and banning malicious content.
Marco Aurélio Ruediger, director of the FGV/DAAP, says that, in political discussions, robots have been used across the party spectrum not only to win followers but also to conduct attacks on opponents and forge artificial discussions.
“They manipulate debates, create and disseminate false news, and influence public opinion by posting and replicating messages on a large scale,” Ruediger says.
The fact is in the war against the bad robots, the weapons used so far have not been enough. In addition to newsroom efforts to provide increasingly qualified information to open citizens’ eyes and ears, it takes active action from giants like Facebook, Twitter, and Google to unbalance the enemy.
And it needs to happen fast. The Brazilian election, happening one year from now, is counting on it.