Publish at September 24 2025 Updated September 29 2025
Under the generic concept of Artificial Intelligence lie thousands of competing intelligent agents. In the world of finance alone, not only are stock exchanges besieged by these financial AIs (more than 50% of all transactions), but the number of AI-assisted investment services is growing all the time. Where it gets crazy is when financial AIs start trading with social network AIs, exploiting their ability to amplify a trend and profit from it. In response, social network operators have learned to moderate these activities, and stock exchanges have programmed alarms and switches to curb suspicious fluctuations.
Comparable examples can be found in the world of politics, where a simple piece of news that resonates can have unpredictable effects. In the social sphere, there are countless social movements, both positive and negative, to which social networking algorithms have contributed: the Arab Spring, strange diets, ephemeral fashions, radicalization, and so on.
The renewed originality of what is posted supposedly makes moderation more difficult, whereas in fact these phenomena of amplification and frenzy follow clearly identifiable patterns. Stock markets have developed a technical response by identifying these memorized patterns of behavior and setting response times adapted to the situation. There's no reason why network operators shouldn't do the same, except that they don't want to.
The basic principle among operators is that as long as these phenomena don't affect profits, it's better not to bother. They may know your food preferences, whether you're expecting a child and who you're going to vote for, but identifying and blocking a notorious fraudster or a chronic transmitter of fake news obviously doesn't earn them much.
In the world of artificial intelligence, anthropomorphisms aren't very popular: we prefer "narrow-minded" to "mono-maniacal" and "focused" to "myopic". As myopes only focus at short distances, we can continue to call them myopic until they are able to vary their focus.
Programming a set of balanced and complex objectives is far more difficult than setting a simple goal like "maximize the metric", (money, clicks, engagement, index of some kind), or "analyze the metric and recommend" (sentiment, opinion, correlation, etc), which can include strategies like "exploit loopholes" or "don't consider the credibility of information". We can be sorry about the amorality of machines, but it's more effective to be sorry about the amorality of their operators, and to supervise them.
In the infinite world of imagination and ideas, we can imagine whatever we like, A.I. can just take us further if we want, but when it comes to applications in a real, finite, complex and balanced world, the use of A.I. comes with ethical constraints.
Since A.I. has no specific purpose of its own, and can perform any task assigned to it, there appears to be no other option than to legally regulate its use in the material world, as is the case for drugs or weapons. The danger is not AI itself, but the hubris of the humans who use it in practice, with no concern for the common good or the environment in which they evolve.
The majority of AIs currently in use can be described as "narrow", because they are designed for a precise, highly focused task. They respond to a well-defined utility or function, with no other consideration for the effects caused outside their parameters; they are monomaniacal and myopic at the same time.
The notion of "prudence" seems foreign to artificial intelligence. At the present stage of development, integrating a feedback loop with a goal in order to modulate its effects throughout a real ecosystem is still beyond our capabilities and, above all, beyond the operators' vision of what a balanced world could be. Hence the need for legal brakes and a regulatory framework.
An A.I. has no moral concerns, and because of the contextual nature of moral judgment, it is not possible to program morality; at most, it is possible to program bigotry. Depending on its use, A.I. can support the quality of life in society (e.g. medicine, education), or on the contrary corrupt it (disinformation, polarization). The problem is not A.I., but the human being, who is content with it in an increasingly technological environment, often becoming dependent on it, without being able to go any further.
The fact remains that A.I. is a tool for amplifying our abilities, not a judge, hence the need for an ethical framework. Technological control tools and a legal framework can be put in place, but human supervision remains necessary, not that of a single individual or a subservient group, but that of all those affected, in constant dialogue.
Basically, an ethical education seems the best way, if not the only way, to preserve our humanity in the concrete use of artificial intelligence. Where else but at school could we do this?
Illustration: StockSnap - Pixabay
References
L'intelligence artificielle en finance - Recommandations pour une utilisation responsable - Autorité des Marchés Financiers
https://lautorite.qc.ca/fileadmin/lautorite/grand_public/publications/professionnels/rapport-intelligence-artificielle-finance-fr.pdf
What is artificial intelligence (AI) in finance? Google
https://cloud.google.com/discover/finance-ai?hl=fr
Flash Crash - https://fr.wikipedia.org/wiki/Flash_Crash_de_2010
Artificial Intelligence (AI) and Investment Fraud: Investor Alert - US Securities and exchange commission
https://www.investor.gov/introduction-investing/general-resources/news-alerts/alerts-bulletins/investor-alerts/artificial-intelligence-fraud
Philosophy for thinking about artificial intelligence - Thesis by François Levin
https://theses.hal.science/tel-04974628
AI and me - Philosophy Magazine - https://www.philomag.com/dossiers/lia-et-moi
Philosophy of artificial intelligence
https://fr.wikipedia.org/wiki/Philosophie_de_l%27intelligence_artificielle
L'intelligence artificielle, objet philosophique - Radio France - Podcasts
https://www.radiofrance.fr/franceculture/podcasts/serie-l-ia-l-intelligence-artificielle
Philosophical Foundations of Artificial Intelligence - Laval University - Courses
https://cursus.edu/fr/34280/fondements-philosophiques-de-lintelligence-artificielle
You might also like:
Self-discipline: a better predictor of academic success than IQ
Self-discipline is one of the fundamental conditions for success in distance learning, and we have every interest in seeing students develop it.
Financial education initiatives
The issue of financial education should be given greater prominence in the education sector. It's true that initiatives do exist. However, strategies need to be found to extend them to as many people as possible, so that education becomes ever more complete.
The right to school: rights go hand in hand with duties
We are immersed in the law on a daily basis, but many rules are acquired by reflex or imitation, without any knowledge of their social, legal or conventional basis. At school, we can take advantage of courses or discussions during which rights and duties are put into perspective.
The distorting mirror: when structure prefers narrative to reality
Behind the rhetoric of benevolence, some institutions sacrifice their best people to preserve a certain internal comfort. Girard, Freyd and Dejours shed light on this mechanism: the scapegoat is chosen, the slander validated, the competence denied. The real institutional manipulator is not the one who lies, but the one who chooses to believe it.
