Elections and networks: Facebook's plan against false news

In a meeting in which he participated domain.comthe social network leaders explained that they would use the most advanced control tools for this election. It is a fusion between human badysts, cybersecurity, artificial intelligence and machine learning specialists to detect what they call "coordinated inauthentic behavior", that is to say that it lacks a mbadive amplification of the dissemination of information in order to misinform the population.

The first step is to detect fake accounts. Then, follow-ups of publications and surveys are carried out based on the comments of the users of the post office or certain characteristics of the content. Similarly, some ratings are verified by independent auditors under an agreement signed between Facebook and the media companies called Reverso. When malicious information is detected, its circulation is limited until it almost disappears. In addition, the ability to get economic returns per click is removed.

"In the past, we did not have a broad enough view of conversations on our platform. It was a mistake and we worked very hard during the last two and a half years to fix it. We want to reduce inappropriate content on the platform, "he said. Stephen Israel, Director of Communications for LATAM. Some milestones mark this breaking point on Facebook: the data breach scandal with consulting firm Cambridge Analytical and the US election that placed Donald Trump in the White House.

For this plan, Facebook brings the experience of several complex elections, such as those in Brazil and Mexico, or those from India, where 950 million people participate, and from the European Union, where the content has to be badyzed in 24 languages.

This strategy has led Facebook to invest billions of dollars in personnel and more sophisticated technological tools.. In fact, they pointed out that since the 2016 US elections, where social networks have played a key role, interactions between users and false news have been reduced by 50%. "We understand the issues, it's a very important moment for the country and that's why we want to prevent our instruments, Facebook, Instagram and WhatsApp, from being misused during the campaign," confirmed Israel.

"There will always be malicious players who will want to take advantage of the platform, we have to have a long lead, it's like an arms race." Marcos Tourinho, Manager of Public Policy and Elections.

On the other hand, looking for greater transparency, Facebook will detail with a specific label the content of the political ad, which will also be grouped in an "ad library", freely accessible, with all the publications of each page of the party. We will see in detail how much each group spent and how far the opinion was.

The work that Facebook is facing has a close connection with the National Electoral Chamber, which will signal when it will detect publications that violate the law. "The goal is that there is no manipulation of elections through our platforms," they ratified. In this sense, they badured that until now, there was no evidence that an attempt was made to do so.

As, Julieta Shama, responsible for strategic alliances with Facebook Media, pointed out that misinformation campaigns will be conducted according to an approach based on three points: "Delete, reduce and report". Any content that violates the social network rules will be erased and any other content that defeats the best user experience will be reduced. For example, the category includes "Clickbait"Because publications are called titles that attract readers, their content does not meet certain quality and value requirements. This type of action "reduces the distribution of false news by 80% on average", which is the second axis.

Regarding the third point, that of reporting, Facebook puts forward the application of the "Context button" which is added to each publication, and where questions such as who created the page, if it was merged with others, and what changes were made to the content are detailed.

There will also be measures to Instagram and mainly WhatsApp. In the first case, the network itself has image verification tools. As for the messaging application, some of the complexities are related to the fact that the conversations are encrypted and the content is not checked. Although they are working on a solution, they have indicated on Facebook that implementing the "send back" tag in messages reduces content circulation by 25%, while highlighting the importance of new permissions. required for the inclusion of a group user