1. Introduction

After exploring the “transformative” impact of AI, I continue my wanderings on generative artificial intelligence (hereinafter GAI) because what is at stake goes beyond the simple use of an LLM in one’s professional activity. Commentators on the subject regularly write about how AI will change our professions. This is a “micro” or verticalized analysis that focuses on a sector, an activity, or even a profession. Although it is difficult to consider the subject in its entirety given how complex things can be, examining the question from a broader angle should also be considered, because if GAI can change, transform, or perhaps even replace certain professions, these changes will inexorably have an impact on “our society.” GAI can impact our professions, but it also has the potential to modify the structure of our society in its “system” aspect.

2. The Society-System

This more “macro” reflection requires making abstractions, taking a step back, and also starting from hypotheses that could be contradicted or contested. This is above all a personal exercise that I wanted to undertake.

Therefore, if we consider that work is not limited to the simple execution of tasks but that it contributes to and participates in maintaining a system, we broaden the perspective of the impact that GAI could have on the system.

This society-system functions thanks to the interconnections of individuals who work together through interconnected flows. These flows are of various natures and touch every corner of our society-system. We naturally think of economic flows, but the interconnections between individuals create flows at the cultural, social, educational levels, etc.

This society-system is a dynamic network nourished by this interconnection of flows, and isolating one aspect of this society-system does not allow us to approach the criticality of the changes that GAI may bring.

2.1. The Individual Is Part of the Society-System

Although our (ultra) capitalist society-system steers us toward a productivist logic, considering that we all work solely to produce seems reductive and not representative of the entire population.

I will take two examples to illustrate these alternatives:

1. The increasingly pervasive search for “meaning” among Generation Z compels us to rethink work (on the subject, see the work of Julia de Funes, who offers an interesting analysis, for example). This search for meaning is not just a whim of young people; it is also present among more experienced individuals who decide to leave everything behind to train for a new profession. Career changes can also be an illustration of and a response to this search for meaning.

2. Degrowth and “low-tech” theories also carry a different vision of our system and aim to be a response to certain excesses engendered by the dominant ideology implemented since the end of World War II. The market logics integrated into our system are being challenged, and their objectives are being questioned.

I do not intend to do anthropology or sociology here, but we must acknowledge that our society is composed of heterogeneous individuals, each working in their own way toward the system that is our “society.” The search for consensus among these diverse aims, notably through politics, allows us to find the balance conducive to everyone’s well-being, sometimes with a utilitarian view influenced by the liberal system established in much of the Western world for nearly 70 years now.

We observe that each individual, notably through their profession, participates in maintaining this system. This society-system, while primarily based on a capitalist and liberal logic, allows other currents of thought to coexist. In this democracy, ultimately, we work in connections, through human interactions.

Each person therefore has their role to play in these interactions, and we all form part of this system we call “society,” like a cog within a great mechanism. “It takes all kinds to make a world,” says the saying. And in a certain sense, it is true. To borrow an expression from Habermas in his article “Labor and Interaction”:

The work of each person is, with regard to its content, a general work for the needs of all.

Let us illustrate this point: if I take a case I know well, the lawyer does not merely practice law in a productivist logic; they also participate, sometimes unconsciously and/or involuntarily, in broader aims by enabling a democratic society to function. The CJEU recalls this in a ruling from September 2024 in these terms:

Lawyers are entrusted with a fundamental mission in a democratic society, namely the defense of litigants.

I am not writing here that the lawyer is an essential element of the system, but they are a link in it. And the same applies to every profession. The street sweeper, the doctor, the carpenter, and the industrialist – each participates, at their own scale, in the functioning of the society-system. I can extend the reflection to every individual who holds, in a certain way, a function within the system. This hypothesis is a form of “structural Darwinism” that allows the balance of the society-system in its functioning.

If this balance is disturbed (which some consider already precarious or simply nonexistent, but that is not the subject here), the system will adapt to find a semblance of balance or a new equilibrium. Our capacity for adaptation is well proven, and past technological evolutions have been integrated into said system, offering in the process a more or less successful demonstration of the resilience of our society-system.

3. The Evolution of the Society-System

I am stating the obvious by writing that our system will (once again) evolve under the impetus of GAI. The extent of this evolution is unknown, and asking the question of the direction in which we wish to go becomes, for some, a duty, even a responsibility. A sort of moral debt toward future generations that responds to the observations currently made regarding choices made by previous generations.

The picture painted above is not complete and can be nuanced. However, the proposals made about the evolution of the society-system often revolve around the same question:

In what society do we want to live?

At first glance, it may make sense to ask this question individually (or collectively). Nevertheless, the answer to these questions involves many considerations that seem uncertain and debatable to me.

This evolution is currently observed at the level of employment, the first “thing” influenced by GAI in the society-system. I will first examine the impact of GAI on professions and then consider the situation of young people more specifically.

3.1. Jobs Are (Still) Being Made Obsolete by Technology

The evolution set in motion by GAI seems inexorable, and the impact on professions appears even more definitive.

I like to cite Alfred Sauvy:

Don’t complain that technical progress destroys jobs; that’s what it’s for.

This situation results from the capitalist logic implemented since the 1960s (or even earlier if we refer to Marx’s writings). Some advocate alternative solutions, and I do not intend to judge the relevance and/or adequacy of the current system and the proposed solutions at this stage.

However, I am fairly convinced that our system operates in pursuit of efficiency in all things. This thesis, defended notably by Jacques Ellul from the 1950s onward, continues to appeal to me for its simplicity but above all for its stunningly current relevance.

We seek, quite spontaneously, to optimize jobs that are insufficiently or not efficient enough in the society-system. To return to the oft-cited example, water carriers and lamplighters were replaced by technology, and the system was reshaped by the arrival of new technologies. These technologies were not necessarily implemented to replace these professions, but the disappearance of these professions is one of the consequences of their arrival. Collateral damage, some would say; creative destruction, according to others – the reality is that pipes and irrigation made water carriers obsolete. Electricity and the light bulb did the same for lamplighters.

In 2025, the ability to use GAI in its agentic form multiplies the field of possibilities and allows for a new (re)modeling of the society-system. If tomorrow you have the ability to control several AI agents that have been programmed to perform certain specific tasks, you reduce, in effect, the number of people needed to carry out all these tasks. But if an AI agent is capable of performing these tasks, what will you be needed for? Your client could just as well use this AI agent, or create their own, and your job would become obsolete (see on the subject some reflections following a question I asked).

We can also illustrate this point with the consultancy giant: McKinsey. After announcing spectacular productivity gains in a 2023 report (later contradicted), McKinsey announced in August 2025 the use of 12,000 agents and a workforce reduction of 5,000 people.

I must admit that I remain skeptical of these announcements, but if we take this announcement at face value, 5,000 people saw their jobs considered obsolete. However, this information should be taken with the greatest caution. First, because McKinsey sells AI services. Having predicted productivity gains in 2023, McKinsey must necessarily demonstrate that it was not wrong (although others wrote so afterward) so as not to lose face before its current and future clients. A self-fulfilling prophecy was needed.

Furthermore, the impact of GAI allows some to mobilize a McKinsey consultant by typing on their keyboard and interacting with a conversational interface. The tips and tricks for leveraging consultancy firms’ best practices are countless. Consequently, this announcement perhaps hides a much more “obscure” reality. If we consider that 12,000 agents can replace 5,000 people, it is perhaps also because the need for consultancy on the market has diminished and this sector will experience a form of obsolescence (although I believe that the competence of certain professions will allow them to “resist” the self-diagnosis engendered by these “replacement by GAI” phenomena).

3.1.1. The Absolute Pursuit of Efficiency Is a Lure

I very much appreciated the analysis proposed by Tariq Krim on writing:

AI plays an ambiguous role. On one hand, it promises us to “write better,” faster, more clearly, more elegantly. On the other, it reinforces the idea that our raw thought is insufficient. Behind each text generation, there is an implicit negation: we are not enough.

We are therefore once again in a race for efficiency. A perpetual desire to do better, to do more, and ultimately, this is perhaps where the problem lies.

GAI acts here as a mirror. It has, at least, the merit of reflecting back to us our own inconsistencies and weaknesses. If perceived with lucidity and humility, this might be the “indisputable” gain of this technology.

But let us not get ahead of ourselves, because if GAI is a mirror, it is certainly a distorting one. At the beginning of my reflections, I was among those who considered that technology was neutral and that it was the use made of it that defined its impact. The hammer metaphor (build or harm) is invoked ad nauseam to try to convince the undecided, to which I belonged.

Through reading and reflecting on these questions, I came to reconsider my view on the subject. Alain Damasio offers an interesting synthesis of the reasons why technology would not be neutral, which can be summarized as follows:

1. Technological innovation depends on research, which itself depends on funding: there are therefore choices, directions taken in the allocation of this funding. Generally, the technologies we use result from research related to military use. The funding to develop technologies is therefore already oriented.
2. Downstream, a technology induces a multitude of changes, sometimes unconsciously. Alain Damasio takes the example of the refrigerator and its impact on our diet and our ability to consume different foods. We can also take smartphones or video conferencing, which have changed our modes of functioning, both private and professional.
3. A new relationship with the world: the arrival of technologies changes our society-system. We are free not to use technology, but it becomes impossible to act as if it did not exist. Damasio takes the example of cars, which have profoundly transformed our cities.

The absence of neutrality of technological tools leads us inexorably to a dead end. Thinking that by layering the latest technological novelties one on top of another we will solve the problems we face is a chimera, and as Gilles Crouch writes:

Technologies often seem to solve problems, but in reality, they tend mostly to reorganize human social relationships and power dynamics.

Technological evolution and capitalist logic will always create new needs and/or new problems that we will have to fill with new tools. True or not, Henry Ford said that if he had asked his customers what they wanted, they would have asked for a faster horse and not an automobile. Steve Jobs said that often, customers do not know what they need until you show them.

This logic is integrated into our Western society-system. This perpetual race will ultimately never lead to a stable and lasting solution. Of course, if you and/or society decide that this is what you/we need, then the reflection can stop there.

If we take GAI itself to illustrate the point, we observe that the evolution of language models rests on a purely commercial logic driven by AI system providers or general-purpose AI models: always more. A faster model, a more precise one, one that can do more things. It is an assumed “flight forward.” Some attempt other approaches (more economical or rational models), but it is always the logic of efficiency that prevails because our society, our system, has been conditioned for this.

Conversely, if you perceive the problem linked to this perpetual pursuit, let us continue. And I do not want to appear reluctant toward technologies, because I am fond of them, but it is this appetite that makes me aware of the excesses we face.

3.2. The Impact of AI on Young People and Employment

A Stanford and MIT report (Brynjolfsson, Chandar, Chen, 2025) reveals that young workers aged 22 to 25 would be among the most affected by this wave of automation, particularly in exposed professions such as software development or customer support. https://digitaleconomy.stanford.edu/wp-content/uploads/2025/08/Canaries_BrynjolfssonChandarChen.pdf?trk=comments_comments-list_comment-text

Behind this initial analysis, which is not very representative overall, lies a fundamental question: what work ethic must we build in the era of GAI? Is it still a simple economic problem or an issue of social justice?

What seems to be emerging is that the age pyramid within companies is being chipped away at its base. Junior profiles generally tasked with performing low-value tasks are being substituted by senior profiles’ applied use of GAI.

It is often said that GAI is like an intern, and this study seems to confirm this view.

I believe this trend will continue to be confirmed and that the result of this first “small” survey will also be observed on a larger scale for a simple reason: the senior profile, thanks to their experience and competence, manages to mobilize the relevant concepts for interacting with the machine much more quickly than the junior profile who fumbles.

The senior profile knows what they want, and with an average level of training, they arrive more or less quickly at what they desire. Training junior profiles becomes “wasted” or superficial time that could be used to “pilot” the machine for production.

Of course, this almost apocalyptic vision will not hold in the long run, and we will, hopefully as quickly as possible, have an awareness of the societal dead end toward which we are heading if we continue on this path: because sooner or later, senior profiles will grow weary of typing or talking with the machine. We will realize that ultimately, junior profiles will be lacking (I deliberately do not address all the other contributions that multigenerational work enables, and this choice is not intended to minimize its impact).

4. In What Society Do We Want to Live?

This is the question everyone asks, claiming through it to have to create and/or adapt a scale of values that would suddenly become universal.

I am quite dubious about this way of proceeding. As if the awareness of the greatest number would set in motion an inexorable force.

In his book “Un taylorisme augmente. Critique de l’intelligence artificielle,” Juan Sebastian Carbonell cites David Noble:

A war is raging but only one side is armed: this is the summary of the question of technology today.

I think that says it all, and I will try to explain why.

4.1. Who Is This “We”?

The question of choosing the society-system in which we would like to live is often posed by people who approach AI on an almost philosophical level. I do not think this question was formulated by a student who uses AI to do their homework or write a dissertation for school. Nor was it posed by the employee who uses AI to write emails, create reports, or presentations to satisfy their employer’s demands, and even less by people who have never used GAI.

In short, this question is raised by thinkers who reflect, often among themselves, on GAI, if I may say so. Categorizing these people is not pejorative or a form of denigration on my part; it is an observation I make having participated in dozens of events, discussions, and trainings on GAI since 2023. This does not mean that other individuals cannot share these reflections. We observe, moreover, that people who train themselves in using GAI very quickly arrive at these considerations because they perceive the underlying stakes of the technology and also grasp the scope of the transformation I am trying to describe. However, the audience targeted by these reflections is infinitesimal compared to the population able to use GAI.

Consequently, if “we” must decide in what society-system we want to live, the “we” would concern, today, only a very small sample of the society-system and the population that composes it. This sample is therefore not particularly representative, and setting in motion a reflection of such importance seems to me to require being brought before the greatest number, even though as users/consumers, it is not the primary question we ask ourselves when we use GAI to accomplish a more or less significant part of our work. I think this absence of reflection is dictated, again, by the productivist logic I describe above. The thinkers therefore raise an interesting, even existential question according to some, but one that seems quite disconnected from the concerns of individuals within the society-system. Let us try, however, to understand why this question is being asked.

4.2. Why Such a Question?

The “why” is important because the reasons can be very different from one person to another: all of this falls within a sort of philosophical, sociological, and political questioning. Some challenge GAI because they do not perceive its benefits in the same way as others. There are substantiated claims but also, sometimes, reflexive rejections.

Others consider that GAI has real benefits to offer. Personally, I believe in the transformative power of GAI on our system-society, although it is necessary not to rush headlong blinded by the promises of big tech. I moreover consider that the reflection on the future of our society-system is a consequence of the (legitimate) fears that certain thinkers have regarding GAI. It is because they precisely perceive the advantages but also the risks and dangers of this technology that they seek to find the balance to avoid the catastrophic scenarios imagined by some.

This aspect seems fundamental to me. Today, beyond the fascination with GAI, we must also think about this new society-system to understand what will make us valuable, because it is not enough to consider GAI as a simple feature. GAI can have a profound impact on our society-system because it will not take our jobs; it will make them obsolete. And this is neither a goal nor a sinister design. It is a consequence already observed, an evolution of the system. We should not necessarily see in it a manifest will to replace humans but simply a manifestation of our inability to solve our own problems. This permanent pursuit of efficiency, once again, is merely a pretext for humans to try to solve problems they have always had.

As mentioned here, independently of GAI, we could also ask whether the unreflective power that tech giants hold, thanks to the business model of the technology economy, is “beneficial” for society as a whole. This reflection is moreover underlying to the initial question and allows, I think, to reinforce the importance of this reflection.

As Carbonell notes in his aforementioned work:

“Ethical” AI and regulation do not question the use of AI as such, only certain of its uses. One could say that it primarily aims to make AI acceptable, presenting it as “responsible.” A well-known example is “The Trolley Problem” in the case of autonomous driving. (…) The way “AI ethics” poses this problem eludes fundamental questions: should this type of AI exist? Or is this a problem that should be solved by an AI?

Carbonell also takes the example of the use of artificial intelligence systems in hiring procedures, recalling the numerous abuses that have been observed in the past, notably at Amazon. Here again, according to the author, we avoid asking whether this kind of decision should be automated and delegated to an artificial intelligence.

In this framework, the question regarding the society-system in which we want to live could take this aspect of the reflection into account. Knowing to what extent we want to or must delegate automated decisions to artificial intelligence.

The “why” is therefore, in my view, fundamental because it allows grasping the true stakes of using GAI and will allow providing a representative and/or democratic response, because as Carbonell states: the problem is neither ethical nor technological, it is political.

Let us then consider this question as eminently important; we must define the “what,” the “who,” and based on the “what,” define the “how.”

4.3. What Can “We” Decide?

What can we decide in the face of this question? A direction in which to guide our society-system? A scale of values to respect? Upon digging deeper, we realize that the answer to what we could decide is not so obvious and that it is much more complex than it appears.

If we set aside the content of the decision, it would nonetheless take the form of a plan that would set out the details relating to the development and use of GAI in our society-system. We could frame it either through soft measures (guidelines or recommendations, …) or through strong measures (prohibition, regulation, …).

In 2025, we observe that recommendations and guidelines have already appeared in a plethora of sectors. These soft measures are likely to raise awareness among individuals in certain uses but generally have no other impact on the society-system.

The transformation is underway, and if we hope to dictate another direction, it seems that strong measures are necessary to “correct” course. European regulation is a perfect example: certain practices are prohibited (social scoring or subliminal manipulation), legislation addresses risks to direct AI systems under development and deployment.

However, despite a good awareness of data protection aspects and the risks of GAI hallucinations, this will not prevent an employer from parting ways with a worker if they deem an AI agent more performant and/or less costly than a human worker.

If “we” consider that this behavior is contrary to the established values (from a social and therefore ethical standpoint, for example), then it will be necessary to put in place measures to curb this behavior. These measures can take several forms (prohibition, sanction, taxation, worker protection against dismissal for “replacement by a machine,” etc.). The objective here is not to find “what” to do but simply to observe that what “we” could decide for our society-system will need to be accompanied by sanctions to achieve the effectiveness of this plan that “we” would have decided for our society-system.

4.4. Who Can Really Decide?

As interesting as the question is and as enriching as the debate on the subject can be, reflection without action is useless. If we take the hypothesis that we must decide in what society-system we want to live and that the plan “we” would have decided must be accompanied by sanctions to ensure its effectiveness, given the spectrum of measures to be implemented, then we should entrust this to those who decide the rules of the game: the members of the legislative power.

This delegation is not new in our society-system, which has integrated a mechanism for delegating individual power to a representative assembly that has the prerogative to manage the State, among other things. Everyone has been able to vote for representatives who have the prerogative to make the State function, itself part of the society-system.

It would therefore be necessary to achieve a global, or at the very least very significant, adhesion to what “we” want as a society-system in order to achieve the effectiveness of the plan.

If we must leave it to partisan politics to dictate our future on the subject, political quarrels will not, in my estimation, lead to a solution consistent with the plan that “we” would have decided.

Moreover, the functioning of our democracies does not allow us to vote “concretely” on a plan but to elect representatives on the basis of a program that must be implemented and whose failure to implement is only minimally sanctioned (essentially through the ballot box).

Habermas reminds us:

One votes for representatives and not for general principles destined to guide future decisions.

The accountability of politicians is therefore important, and this is obviously the fundamental issue: how to get the plan through rather than leaving it to political representatives who are discovering, like us, this transformation.

On this point, Habermas had, once again, a luminous consideration:

The human species has challenged itself not only to provoke the social destiny that is its own but also to learn to master it. It is not possible to meet this challenge posed by technology with the resources of technology alone.

According to the German philosopher:

It is rather a matter of engaging a discussion, leading to political consequences, that rationally and bindingly relates the potential that society has in terms of knowledge with our practical knowledge and will.

What we would need, then, is to find a consensus shared by the greatest number, built through exchange and discussion. Utopian perhaps, but it is the objective to achieve in order to succeed in this evolution of the society-system.

5. Critical Examination of a Plan for Our Society-System

It would therefore only be through discussions that we would manage to develop this plan. If so, the solution seems ultimately quite simple.

The organization of great debates, referendums, or any other mechanism of direct democracy would allow deciding on the general principles destined to guide future decisions, to use Habermas’s terms. A prerequisite for these popular consultations would then be to inform the public in a coordinated manner about the situation. But inform them how? To say what? Who will decide the topics to address? We are going in circles, in fact.

In reality, I think this entire discussion is an illusion of a solution whose only purpose is to provide a regulative principle. In other words, it is observing a problem and evoking an out-of-reach solution like a Kantian ideal: an idea that is often considered a real object when it is merely a regulative idea.

Moreover, this question of “in what society do we want to live” will not solve the observed problems. The definition of a grand plan to redirect society is moreover an inadequate means given the structural and cyclical situation of our society-system and will engender harmful consequences.

5.1. A Prerequisite: Making Society or the Search for a Lost Paradise?

This reflection leads me to type this funny title, inadvertently referencing a naturist documentary from the late 1990s. Do not see any mischief in it, only a light interlude in this reflection. Let us now return to the matter at hand.

Before deciding on the contours of the society-system in which we wish to live, we will first need to make society.

And in this context, “making society” is not just living together or working together. It is participating in the society-system in which rules and an organization exist. By making society, we oblige each other to live in (more or less great) harmony. Work becomes a way of creating bonds, chains of interdependencies that help us make society. The fact that a profession has meaning is not just an individual or isolated perception; it is a whole. It is because the society-system values it and gives it a certain legitimacy.

Therefore, if a GAI can perform tasks and therefore professions, how will our society-system integrate this novelty? Can GAI also make society with us, as we try to do?

We can clearly see on the subject of GAI that deciding in what society-system we want to live, in such a polarized environment, is manifestly very complex. The consensus we seek to find seems quite illusory at this stage of our reflections and our societal evolutions.

As described above, the broader reflection on our society-system invites each person to position themselves on the society-system in which they would like to live: from a society governed by efficiency and therefore with AI everywhere, to the opposite end of the spectrum (which I leave you to define according to your convictions and reflections).

However, it is perhaps by redefining this new society-system, together, that we will manage to make society. The stake is therefore, if this hypothesis is validated, even more important.

5.2. The Social Inadequacy of a Global Plan

In a certain way, if we manage to coordinate to establish these regulations, they would be comparable to the grand plan mentioned by F. Hayek in his book “The Road to Serfdom.” This/these grand plan(s) found by consensus contain(s) a weakness described by Hayek that seems interesting to examine in order to better perceive the regulative ideal that this idea engenders.

If we want to identify the guidelines for defining this “grand plan for our society-system,” we will certainly be inspired by the pursuit of a “general interest” or a “common good” in determining what is tolerable or not. This objective is a trap into which we should not fall because it leads to disastrous consequences according to Hayek, as he states:

The welfare of a people, like the happiness of a man, depends on a great number of things that can be provided in an infinite variety of combinations. It cannot be defined as a single end, but as a hierarchy of ends, a complete scale of values where each need of each individual receives its place. Directing all our activities according to a single plan presupposes that each of our needs is placed at its rank in an order of values that must be complete enough to allow choosing between all the directions from which the planner must choose. This presupposes, in short, the existence of a complete ethical code in which all human values are placed in their legitimate position.

There will therefore be, objectively, a form of denial of certain interests in implementing a grand plan. Certain interests will have to be privileged without objective reason, and a permanent arbitration of values will be necessary to define the contours of what we want.

Moreover, due to the necessity of resorting to sanctions and therefore regulation, the global character of this regulation must be the objective.

To return to Hayek’s point:

It is easy to convince the people of a given country to make a sacrifice for “their” iron industry or for “their” agriculture so that no one in the country is forced to lower their standard of living. But one need only imagine the problems raised by economic planning encompassing a region like Western Europe to understand that there is no moral basis for such an enterprise.

According to Hayek again, the implementation of a plan, on an international scale even more than on a national one, is merely the application of brute force: a small group imposes on everyone else a standard of living and a work plan that it considers just.

5.3. Is This Therefore the Society We Want?

I have tried to push the reasoning to excess. The directions taken in my reflection are certainly contestable, and I am happy to debate them in order to better understand the reflections of those who think I am mistaken.

But I do not intend to escape at this stage of the reflection because the solution proposed by Hayek in the face of all this is:

What is needed is an international political authority which, without having the power to prescribe to people what they must do, is capable of preventing them from harming others.

He advocates for respect for the individual and therefore for individualism (the opposite of socialism). And individualism should not be taken as a form of egoism but rather as the act of “respecting the individual as such, recognizing that their opinions and tastes belong only to them, within their sphere, however narrowly circumscribed, is to believe that it is desirable for men to develop their individual gifts and tendencies.”

The proposed solution ultimately amounts to prescribing one thing to people: preventing them from harming others. In this framework, transposing such a solution seems difficult in the era of GAI, unless one considers that the fundamental freedoms of the individual enshrined in international texts are protected and respected in an even stronger and more important manner. A project of equally significant scope.