On Saturday November 7, after more than three long days of counting the votes, Joe Biden was declared the winner of the US presidential election. Based on the votes obtained, the American media considered that his lead over incumbent President Donald Trump was irreversible.
The ballot and the days that followed were peppered with a lot of false information and part of the American public is still convinced, against all evidence, that the ballot was ” Fly “. How to explain this influx of disinformation, despite the precautions of the platforms? Camille François, specialist in disinformation on social networks, responsible for innovation within the company Graphika and member of the Election integrity partnership (EIP) – a group for studying disinformation during the presidential election bringing together researchers and experts – answered our questions.
Page Contents
How did the spread of disinformation during the election and the days after it evolve?
Volume is difficult to measure, as the platforms have been very active in moderating content. That said, the structure of disinformation has had fairly clear phases. On election day there were a lot of isolated incidents, with internet users saying “This is what happens in the polling stations”, “I received the wrong bulletin”, “I saw someone with a felt-tip pen, is that allowed in the polling stations? “. Many of these incidents were false, with the exception of two or three, which received wide coverage, particularly in Philadelphia, where there was a great deal of misinformation on election day.
These disparate incidents were then organized into “narratives” [une histoire qui va agréger des éléments disparates et leur donner un sens commun]. The “Sharpiegate” [une théorie selon laquelle il a été donné aux électeurs républicains des stylos-feutres rendant leurs bulletins invalides] is a pretty clear example. It’s a complicated moment, because even though the incidents have been proven false, the force of the narrative continues to carry disinformation.
These narratives then gathered in movements and mobilizations. On Wednesday evening, conservatives rallied around the hashtag #stopthesteal [« arrêtez le vol »]. A Facebook page was created by political figures close to Donald Trump and the radical right, and very quickly had enormous success; it was deleted by Facebook in the morning. This page took hold of those narratives and began to organize events. This is the phase where disinformation takes to the streets and generates calls for violence. This moment when misinformation started to impact offline life is a turning point for me.
Today, an ecosystem of alternative media still denies the election result, continues to say the election is not over, and amplifies these disinformation narratives.
We have the impression of something that is growing without being based on any facts. Why does it keep getting bigger?
This is due to several factors, for example psychological (such as cognitive biases), or relating to the structure of the American media and political landscape. We see that even when we demonstrate that the videos are fake or that the events did not take place, we continue to see virality processes. In 2020, one of the novelties is the role of alternative networks, like Gab or Parler, invested by hard right-wing figures who complain about the censorship of Facebook, Google and Twitter because of their much higher standards of moderation. high.
What explains the popularity of these theories, despite the fact-checkers?
From a data point of view, we can see that there are a handful of political influencers, in particular “alt-right” [d’extrême droite], which play an essential role in the dissemination of these theories. This group takes the narratives and transforms them into a political movement. It is this group that transforms various conspiracy theories into a mobilization movement around the slogan “Stop the Steal”.
Can we tell how many Americans have been exposed to disinformation?
Disinformation has taken a leading role in this electoral cycle. Between the Americans who saw the disinformation incidents, those who heard elected officials talk about it and those who consulted the media fact-checkers, there was immense exposure to disinformation, which does not mean everyone believed her!
Is electoral disinformation using video passing through the net of moderation?
Yes, but I think within the videos there are certain types of content and platforms that are more difficult to study. Many researchers have pointed out the difficulty in studying disinformation on YouTube. But TikTok, for example, poses sometimes even greater difficulties, because there is no external access to the data, unlike YouTube. And of course, the “livestream” format, of live video, also poses significant issues of moderation during an election period.
There was a lot of talk in the months leading up to the election about Facebook and Twitter, and much less about YouTube. Have we seen more disinformation?
The conversation about disinformation in recent years has focused a lot on Facebook and Twitter, and YouTube issues have taken a back seat. This is regrettable because YouTube – and in general Google [qui possède YouTube], because there are also the advertisements and the home page which can participate in the rapid dissemination of disinformation – is affected like other platforms by these topics.
Is YouTube doing enough in the fight against disinformation?
Several dimensions must be taken into account to answer this question. Does the platform have specific rules against the different content subtypes that make up the so-called “disinformation” set? The “disinformation” category is often quite broad, and can contain electoral “fake news”, such as fake Russian accounts, the two having little in common. For example, the comparative study we conducted within the EIP showed that Google had fewer rules on certain subtypes of electoral disinformation.
The second question is whether the rules are applied consistently and effectively. Then, are there teams in place, whose job it is to detect disinformation? It is often thought that this can be done by artificial intelligence. This is quite wrong in general, and very wrong in the context of an election. Finally, there is the issue of transparency. How do the platforms communicate about their processes? On these issues of transparency, there are a lot of differences between the platforms.
If we look at these criteria, can we say that YouTube is lagging behind its competitors?
On these issues, Google has not positioned itself as a leader in the industry, and still has a lot of good practices to adopt.
Did the large platforms successfully pass this very delicate period?
It is too early to tell: the period of election uncertainty continues in the United States. In general, their mobilization is clear, even if small platforms, like TikTok or Pinterest, have also mobilized. The networks were completely caught off guard in 2016, especially with regard to Russian electoral interference, and have come under a lot of pressure from elected American officials but also from civil society to do better. It is thanks to this pressure that the giants of the Web have deployed exceptional means. The question now arises as to whether these efforts will be continued, not only over time but also beyond American borders.
Two days after the election results, despite action on both sides, there are still millions of Americans who believe the election was stolen from them. How to explain it? What could the platforms have done?
It would be a trap to think that all issues of disinformation are issues related to the Internet. There are deep political issues: even if the platforms and the entire media ecosystem have an important role in its dissemination, it cannot be isolated from other political and social factors specific to the United States. .
Will this narrative of a “stolen” election have a long-term impact in the United States?
Some of the people who think the election was stolen also have high hopes for the American legal system. These theories can fall like a breath, but can also fuel conspiracy theories very present in the United States. This narrative resonates with conspiratorial communities, well established today in the American political conversation, which do not trust the “deep state”.