Bots boosted half of Bolsonaro's supportive tweets at the start of the campaign

Bots boosted half of Bolsonaro’s supportive tweets at the start of the campaign

Supporters of President Jair Bolsonaro (PL) and former President Luiz Inácio Lula da Silva used bots to promote their candidates on Twitter on the first day of the campaign. Automation use was higher among Bolsonaristas: 50% of retweets, versus 25% among PT supporters. This practice is not permitted by law.

The behavior was determined by the Pegabot tool, an algorithm developed by ITS-Rio (Rio de Janeiro Institute of Technology and Society), which was formed by researchers from universities such as UERJ (Rio de Janeiro State University), PUC-Rio (Pontifical Catholic University of Rio) and MIT (Pontifical Catholic University of Rio). Massachusetts Institute of Technology) Media Lab, who shared the survey exclusively with UOL.

The campaign officially began on August 16. At the time, Pegabot on Twitter analyzed the behavior of the tags highlighted that day: #Vote22Bolsonaro and #LulaPresident13.

“These hashtags have amassed a relevant volume of stock in a short period of time,” ITS-Rio says of the bots.

How do robots work and what did Pegabot discover? These bots work by automatically and sequentially retweeting posts from real people using a specific hashtag. This makes it appear in more posts, which increases its performance on Twitter and is among the most talked about topics on the social network.

The institute explains that on average, the same user mentioned up to four times as many hashtags #Vote22Bolsonaro and #LulaPresident13. For the survey, users who submitted at least five times were selected, and “we analyzed the presence of automated behavior among them using Pegabot.”

On the first day of the campaign, Bolsonaristas were responsible for 53.5% of 125,300 tweets analyzed, versus 46.5% of Labor members. Part of the advantage is due to the role of bots, a larger role among the president’s supporters.

What was the result? The conclusion is that 49.8% of the 18,926 Bolsonarista users analyzed “provided a probability of automated behavior with a value equal to or greater than 70%”, which represents “with 43,908 tweets and retweets shared, 65.4% of the total volume of tweets collected [67,1 mil]”.

Among PT members, 24.7% of 22,286 assessed users demonstrated automated behavior, sharing 26,816 tweets and tweets, or 46% of the total collected. [58,2 mil]”.

Some users have even shared tweets and tweets more than 200 times.”
Institute of Technology and Society of Rio de Janeiro

One Bolsonaro bot reached 434 posts, while the PT bot did the same 359 times.

How does Pegabot find bots on Twitter? Karina Santos, who is in charge of ITS-Rio’s Department of Democracy and Media, says the institute monitors hashtags daily, “and identifies tags that appear in a potentially suspicious manner,” as was the case with #Vote22Bolsonaro and #LulaPresident13.

“The speed of participation and the type of interaction around a topic indicate the potential for inauthentic behavior,” he says. “The algorithm analyzes the characteristics of the profile, the type of post on the platform, the language used in the account, and the network of followers around the user.”

Why is using bots dangerous? The expert warns that in social networks, artificial intelligence can be used without transparency “to amplify specific discourse, build and spread narratives, create content bubbles and spread misinformation.”

We see a powerful use of bots to disseminate information with the aim of deceiving public opinion and falsifying a sense of importance and support. They are used to make content appear to have been commented on by everyone without it being, to discredit certain characters or praise others.”
Karina Santos, from ITS-Rio

a What is the opinion of the law? The electoral law allows content promotion on social networks only if the author has paid the platform, such as Twitter, to disseminate the information. In this case, the “Sponsored Content” notification appears.

Public law attorney Roberto Becceli, who cites Law 9.504/97, amended in 2017, says the use of robots is prohibited. One excerpt states that “it is not permitted to broadcast electoral content by registering an internet application user with identity fraud intent.”

And in another article, it states that “the use of enhanced content and digital tools not provided by the Internet Application Provider is prohibited.”

What if the person responsible for the robot is arrested? The Supreme Electoral Tribunal (TSE) only intervenes if it receives a formal complaint “from any citizen, including opponents,” explains Bekele, who studies the relationship between the constitution and digital media.

The law provides for a fine of R$5,000 to R$30,000 for the author, if specified, as well as for the beneficiary candidate “on proof of prior knowledge”. The social network will only be fined if it fails to comply with a court order to remove the content from the air.

“As far as we can assume that the initiative came from users without the knowledge of the campaign, the use of bots violates the logic of electoral expense control,” says the attorney. “After all, it is an unofficial contribution to the beneficiary candidate’s campaign, most likely from an illicit source.”

What do the campaign candidates say? After that, the campaigns of the candidates had not responded until the publication of this report.

According to Twitter, third-party tools that “access public account information” collect “very limited data for this analysis.”

The platform states that “the Supreme Electoral Tribunal has a direct and permanent channel of communication with Twitter to discuss initiatives, best practices, and actions needed to protect the integrity of conversations around elections.”

“This is in addition to the regular and frequent enforcement of our rules, including those related to combating spam and platform manipulation, and the increasingly proactive detection of potential violations of it,” he says.

According to data recently released by the platform, out of the 238 million active profiles on the network, 5% are triggered automatically. Billionaire Elon Musk questioned the data, who decided to cancel the platform’s purchase, alleging Twitter’s lack of transparency about fake profiles.

Karina advocates for regulation in the style of “follow the money,” an expression in English that means following the money of those who fund these coordinated campaigns “because Brazil’s market for disinformation is very well regulated.”

It concludes that “in an electoral scenario, this use could have a direct impact on a voter’s voting decision, which is deeply concerning.”


#Bots #boosted #Bolsonaros #supportive #tweets #start #campaign

Leave a Comment

Your email address will not be published.