The book. At first glance, the reader will probably think that he has already read and heard these lamentationss on the harmful effect of social networks on public debate and on the risks of influence of opinions by certain interest groups. Especially since with the war in Ukraine and the French presidential campaign, the time is conducive to this kind of suspicion.
But if in his book, Toxic Datathe mathematics associate David Chavalarias indeed offers one more criticism against the Gafam, Russian or Chinese interference, his analysis is quite original in several respects.
The central color illustrations, which probably few readers have already seen, already testify to a different view of these themes. We see clouds of dots in heaps, or else scattered, connected to each other more or less strongly. This is what the author, a mathematician by training, calls a “macroscope”: a device for studying on a large scale what is happening on a social network like Twitter and which, by collecting and studying millions of tweets, to see where a piece of information comes from, how a keyword circulates, how political communities are formed, and to see them burst, recompose or come closer to another… As a specialist in complex systems, whose social networks are part of it, he defends the idea that to better correct the negative effects of these digital ogres on opinions, it is first necessary to have the tools to “see” them.
A “polytoscope”
Several examples show the interest of such an approach. This “politoscope”, as it has been baptized, has shed light on the fact that the term “Islamo-leftism”, taken up by ministers, is indeed a concept that comes from the extreme right. It also helps to identify which camps are spreading the most false information. It documents the emergence of the antivax and antipass sanitary community.
The social behaviors of echo chamber, filter bubbles, polarization,astroturfing, etc., are of course present but seen in the light of network science and cognitive science, which makes it possible to better understand them. One of these results, taken from a mathematical theorem from 1976, is even quite depressing: the stronger the social interactions, the more the collective behavior is unpredictable in the eyes of individuals (making situations unstable), but the more the entity who has the information on these interactions can predict them…
Very concerned about these influences on the health of our democracies and aware that choices are sometimes made by a number of votes that can be easily reached by influencers, the author outlines 18 proposals to avoid disruptions. Some are common sense: check your sources, know how to disconnect… Others are difficult to achieve, such as ” to watch “ algorithms, “giving data back to the people”… Still others are more unexpected, such as the use of majority judgment for elections, or “Preferendums”, which make it possible to qualify the content of the votes and avoid binary judgments and the inevitable polarization.
You have 1.86% of this article left to read. The following is for subscribers only.