Publish at March 27 2024 Updated April 29 2026
Giuliano da Empoli's excellent book "The Engineers of Chaos" offers an analysis of the direct consequences of algorithmic capabilities in the manipulation and orientation of our reactions.
The formula is simple: millions of comments and news items are published on the Internet every day; those that receive the most reactions are deconstructed, analyzed, distilled into increasingly effective concentrates and then improved, promoted, amplified as data is accumulated on the participants.
In the space of a few years, perceptive analysts manage to develop not a coherent political program, but a force that can be directed at will, amplified or changed direction as long as it finds fuel to feed it: frustration, anger and indignation. There's no need to worry about accuracy, truth or even verisimilitude; only reaction really matters.
These analysts are in a position to control the phenomenon, and they offer their services to the highest bidder, be it Trump, Bolsonaro, Salvini and many others, who have no use for the representative institutions of a power and a system described as unequal and elitist, elements they exploit to their advantage by amplifying it. Cambridge Analytica is just one anecdote in the development of this industry.
Guiding an algorithm seems quite nebulous to most people. Here's an example: "At every occurrence of someone in a hat and briefcase, you note the location and temperature." Today, with artificial intelligence, we can go much further and take note of the entire context, structured or unstructured data.
When enough data is accumulated, an impressionistic picture emerges, and we can guide actions based on this picture, such as broadcasting useful messages to people with hats and briefcases, or offering them services, activities or products at the right time.
This is what social networks and search engines do. They don't worry about potential biases, since their data is derived from our actual activity; they take no responsibility. The fact that almost everyone in a hat and briefcase is white, rich and lives in the same kind of neighborhood doesn't concern them. They're not programmed to reduce inequality, but "to be useful where it pays off". That's the algorithm. They sell their services to the highest bidder; the auction is their modus operandi.
They could assume their responsibilities when they sell their services, but they don't either. They can detect that a user has moved to a different city for the weekend and suggest a local restaurant to suit his or her tastes, but they claim to be incapable of identifying a serial swindler, recognizing a notorious scam or a defamatory fake news story even if it is repeatedly reported to them. They don't have an algorithm for that. The quest for maximum profit is their fundamental programming, not that of the public good. In 2026, the Musk-Altman trial came down to precisely this principle.
Authoritarian regimes, autocracies, mafias, pornographers and other profiteers of human weaknesses would be foolish not to make use of the new possibilities available to them; unscrupulous individuals put these algorithms to work for them, undermining democracies and social gains alike.
These new tools are capable of identifying the manipulation points of entire populations without any ethical consideration or concern for the public good or even for the future. The exacerbation of reactions systematically leads to a destructive frenzy. History is full of this, and even an A.I. can account for it if asked.
Google has changed its slogan from "Don't be evil" to "Do the right thing"; its difficulty is to define what is "right". Meta's "Be bold" has become "Live in the future", without indicating whose vision of the future it is (hers?). Byte Dance (Inspire Creativity, Enrich Life), Alibaba (Global trade starts here), Apple (Think different) present their general orientation, but in all cases these major principles are bitterly debated in the Board of Directors (BoD) and are systematically and without exception subjected to the diktat of profitability for shareholders. The basic algorithm remains short-term profitability. Optimum profitability would consider the well-being of mankind as far as can be predicted, and in parameters other than purely financial ones.
Open AI, the company that put "the well-being of mankind" at the heart of its mission, finally welcomed major investors onto its board of directors, generating an atypical crisis. Will its "not-for-profit" mission stand up to commercial appeal? Will we simply outsource the profitable aspects and preserve a facade of virginity? We can't leave this kind of decision to the goodwill of a few people in control of these tools.
The European Union's responseto A.I. control is in the making, as is that of several democracies. The general principles are well known, and aim to make the operators of these systems accountable at all levels.
In education alone, the implications are numerous, if only in exam marking, career guidance, selection for internships or access to training programs. Justice, trade, infrastructure, employment, the environment, etc. - not a single field is spared.
Sociology, philosophy and most of the social sciences have their work cut out for them, and plenty to study when it comes to the effect of technology on our lives. It's not a question of putting the brakes on progress, but of ensuring that there is progress for all of society, and that it's not defined solely in financial and digital terms.
Illustration: Apophis26 - DepositPhotos
References
The engineers of chaos - Giuliano da Empoli
https://www.decitre.fr/livres/les-ingenieurs-du-chaos-9782073019240.html
Connards malgré nous - How to stay intelligent in the face of emotional manipulation - Olivier Bas
https://www.decitre.fr/livres/connards-malgre-nous-9782100817450.html
Cambridge Analytica - Wikipedia - https://fr.wikipedia.org/wiki/Scandale_Facebook-Cambridge_Analytica
Musk accuses Altman of stealing a charity as courtroom battle begins - Lily Jamali - BBC
https://www.bbc.com/news/articles/cz027nyz529o
AI law - European Union
https://digital-strategy.ec.europa.eu/fr/policies/regulatory-framework-ai
Social impact of generative AI: benefits and threats - Unite AI
https://www.unite.ai/fr/impact-social-des-avantages-et-des-menaces-de-l%E2%80%99IA-g%C3%A9n%C3%A9rative/
Ethics in artificial intelligence: non-negligible impacts - Le bon digital
https://lebondigital.com/lethique-dans-lintelligence-artificielle-des-impacts-non-negligeables/
Meta's new slogans aren't impressing branding experts - or metaverse veterans - TNW - Thomas Macaulay
https://thenextweb.com/news/metas-new-values-arent-impressing-branding-experts-or-metaverse-veterans
Introducing OpenAI - Mission statement
https://openai.com/blog/introducing-openai
OpenAI: "This was no ordinary corporate battle" - Chloé Sondervorst - Radio-Canada
https://ici.radio-canada.ca/nouvelle/2030752/openai-bataille-entreprise-chatgpt-gillian-hadfield
You might also like:
Is it still possible to teach the classics?
Classical works have shaped our societies and created images that continue to haunt our collective memories. And yet, for younger generations, these creations seem passé. Is there a way to teach the classics without raising a wave of sighs?
In Search of the World's Richest Language
Which language is the most lexically rich in the world? Which language can boast the largest vocabulary? Put yourself in the shoes of a linguist and go in search of the richest language!
How do you tell a story? Some authors have looked for the invariants, the ingredients of effective storytelling... They found patterns, principles and tropes present in stories and cultures far apart. There's no magic recipe, but there are plenty of rules to follow, or to shake up!
Trees of knowledge, with many ramifications
Each civilization, each culture considers its trees under a specific angle, this glance teaches us on the reports with the privileged knowledge between morals and freedom...
Between effort and lack of effort, there's a whole list of parameters to consider before judging others.
