Criticisms of current social media platforms are of several natures: some lament at how distracting they are by design, with consequences on both the personal and the professional life. Others, taking a more philosophical stance, argue their design promotes ways of behaving or interacting that takes their users away from many desirable virtues. An angle of attack that recently gained traction and falls into the latter category is to highlight that they are powerful radicalization machines and induce political polarization.

In the early 2010s, the role of Facebook and Twitter in the trigger and organization of the Arab springs gave social media an aura of tools promoting democracy. A couple of years later, culminating with the 2016 Trump election in a backdrop of large scale mis-information and Russian interference, the public perception had reversed.

“The Chaos Machine”, by Max Fisher, tells the story of how the largest social media platforms led to radicalization all around the world, fuelling genocides in Myanmar, anti-refugee attacks in Germany, leading to the rise of QAnon in the USA to name a few. The cause? A mix of financial incentives and hubris. VC funded platforms seek growth at all costs. Their business model is advertising and the goal is the maximization of user engagement - time spent on the platform -ultimately sold to advertisers in a race for users’ attention.

In each instance of radicalization, the causal chain is both hard to establish and complex to explain. It is a mixture of human psychology (our inclinations, impulses and tendencies), pre-existing conflicts or oppositions, of technological determinism: social media are more than simple tools and do change our behavior, and of algorithms with nefarious (mostly unexpected, yet rarely acted upon) emergent behaviors as a result of their goal to maximize engagement.

An early example: attitudes towards vaccination in the US in 2014, studied by Renee DiResta. Users defiant towards vaccination tend to spend more time online seeking and sharing information. As a result, anti vaccine sentiment occupies a much larger fraction of the virtual space than it does in real life. The underlying mechanisms are the ability to reach huge audiences, the removal of friction (it takes one click to share), the different social norms that apply online (people don’t get silenced). On top of that, algorithms do identify and over share posts or invitations to groups that catch the attention the most and lead to further time spent online; these posts tend to be more outrageous, fear inducing, and often factually wrong. When external observers noticing these trends reported their worries to Facebook, they replied with what would become their usual response: the platforms are not to blame, there should be no obstacles to free expression online, they know better how to fix problems the state has been unable to solve so far.

The problem grew bigger as companies relied more heavily on the use of algorithms to recommend content. Such algorithms work by figuring out, by trial and error, what content leads to the maximum engagement for each individual user. The huge amount of personal information the platform gathers allows this recommendation to be extremely personalized and effective.

Whether on Facebook, Twitter or YouTube, these algorithms have been shown to lead users down rabbit holes of radicalization, sometimes identifying and recreating known paths towards radicalization that predate social media. For example the disenfranchised may find relief following the advice of a self-help guru, leading to another one who, en passant, starts to introduce scapegoats that could explain personal issues, ending with hate speech towards a now enemy group to get rid of.

One annoying aspect but also a force of this book is the hammering of the same story unfolding time and time again, with journalists, researchers and whistleblowers exposing the evidence of the role of social media platforms and the refusals of the latter to take any form of serious action to counter it, unless it directly threatens their profits. Its explanation of the economic incentives driving the business of social media platforms and of the psychological and sociological research on how we interact with those is exhaustive and enlightening.

Fisher is not very optimistic about the future: he believes we are only seeing the beginning of the effect of radicalization on the political landscape, to the benefit of once outcast political forces. He sees the US as unable to act against its most successful tech companies. However, he believes regulation might come from Europe - who already shared its impatience with Facebook’s lack of action - and he predicts the fight with the tech giant will be ruthless.