The year 2022 was marked by election seasons around the world. While in Latin America Colombia and Brazil overturned their governments, in the United States mid-term elections renewed the US Congress. As in recent years, social networks were the battleground of these three campaigns, marked by polarization and the dissemination of disinformation narratives.
In this blog, we take a look back at the main challenges of these scenarios, the way in which the platforms acted and the effectiveness of their content policies, which over time have been adjusted to correct and prevent past failures.
Colombia: platforms on the sidelines
The first half of the year Colombia experienced a three-phase electoral season: the legislative and inter-party elections held on March 13, and the two presidential rounds held in May and June, respectively. These three stages allowed some disinformation narratives to run their course and others to become blurred as the events progressed.
In general, the campaign was accompanied by distrust in the electoral system, driven by several political sectors. Among others, theories were spread about the Registraduría's vote counting software and the breaking of the chain of custody of the votes during the legislative elections, which sowed doubts in the lead up to the presidential elections.
As a backstop to digital discussions were the platforms' content standards, which in the wake of the 2016 and 2022 elections in the United States had included measures to control misinformation, such as content labels and penalties for statements that mislead about or delegitimize the process.
Indeed, some publications that circulated during those weeks, including those of high profiles and candidates, appeared to breach the platforms' rules, such as Meta and Twitter's policies regarding content that falsely questions electoral processes or even the rules that prevent the promotion of electoral crimes.
If at the beginning of the campaign there were expectations about the level of interference that the platforms would have to act on problematic content, in the end it became clear that they preferred to stay out of the way and not apply their policies strictly.
Regarding the silence of social media companies in the face of the problematic content that circulated in the three Colombian elections of 2022, there are at least two questions. The first is related to the detection capacity of their systems, which may not be up to the level of detail and context that certain rules require to be applied.
The second one may have more depth and also involves content moderation, its practices and the design of its policies, since it lies in the convenience or not of platforms acting rigorously in electoral contexts, where any measure -a suspension or the removal of a content, for example- may become an element that further aggravates a polarized media.
Brazil: the 2020 that was not
To some extent, the policies that the platforms designed after the 2020 elections in the United States - where the delegitimization of the process and doubts of fraud led to a situation of real instability - would seem to have full correspondence with the Brazilian context, where at times Jair Bolsonaro and some of his allies were willing to push polarization to the limit.
After October 30, when the first round was held and Luiz Inácio Lula da Silva prevailed over Bolsonaro, a wave of nationalist extremism flooded social networks, accompanied by demands for military coups and claims of fraud.
An investigation by the Digital Forensic Research Lab also detected the participation of inauthentic accounts that promoted through different social networks the hashtag #BrazilWasStolen, reminiscent of the #StopTheSteal that spread after the 2020 elections in the United States.
In addition, ads were found - published before the deadline established by Brazilian law to be able to advertise politically - that reproduced speeches in which Bolsonaro called on citizens to march in the streets to defend their lives, a narrative that also resembles the one promoted by Donald Trump after his defeat in 2020, as pointed out by Flora Rebello Arduini, campaigns director of the activist organization SumOfUs.
Brazil's population makes it a very attractive market for social networks, an interest that is also seen in the way platforms treat their elections. For example, over the course of this year, YouTube modified its electoral integrity policies three times to ban content that questioned the legitimacy of the 2014, 2018 and 2022 elections in that country, an attention it shares only with the United States and Germany.
However, as in other contexts, there are problems of detection and consistency in the systems that aim to control problematic content related to elections. In addition to these problems, there is the language barrier, as there has not been sufficient investment to ensure that artificial intelligence and human moderators have the conditions to review content in Portuguese in the same way as in English.
United States: confined disinformation
The mid-term elections, in which the country renewed part of its Congress, were the opportunity to measure the effectiveness of the measures implemented by the platforms from the 2020 presidential elections, which ended with the takeover of the Capitol and the expulsion of Donald Trump from the main social networks.
Although it was expected that problematic content would once again play an important role in the development of the elections, its impact was much less than it had in 2020. According to Alex Stamos, director of the Stanford Internet Observatory, measures such as the suspension of certain voices led to the dissemination of conspiracy theories or disinformation content to be set aside in niche platforms, such as Parler or Truth Social, without the possibility of having a reach like in the past.
In turn, factors such as initiatives by the media, civil society organizations and government entities to promote voter education helped users identify misinformation before the elections.
Although the situation was more encouraging than in other elections held this year, the misinformation indicators demonstrated the failure of the platforms to detect and sanction problematic content in languages other than English. This deficiency of moderation systems, present in the rest of the non-English speaking world, opened a gap in the quality of content to which Latino and Asian communities, among others, were exposed.
The monitoring of the mid-term elections also allowed us to see the growth of alternative platforms such as Telegram and Truth Social. While large social media companies - such as Meta, YouTube and Twitter - have been in the spotlight of the authorities and civil society and have therefore yielded to greater actions of commitment to transparency and the fight against disinformation, some minority platforms have remained on the sidelines of these demands and have not implemented specific rules for electoral contexts.
As we have seen, the community standards of the platforms respond to a large extent to the U.S. context. In the case of Latin America, compliance is not as effectively guaranteed and, when applied, often ignores the differential contexts of each country. Openness to diversity is a first scenario in which platforms are beginning to make inroads. However, there are more concrete actions that can contribute to the construction of a more transparent digital space. For example, education to identify false information, greater support for organizations working on digital public debate, stricter and more targeted regulation and a strengthening of communication channels that allow platforms to meet the needs of their users around the world.