1. Anti-human, All Too Human
We can now anticipate that generative AI will soon be able to produce more knowledge than human beings can. This has raised questions concerning the future of knowledge—questions similar to those that prompted various twentieth-century thinkers to discuss the death of the author, a thread that has been picked up again today in view of generative AI. While Michel Foucault viewed the author as an institutional and social construction, and Roland Barthes heralded the end of authority over meaning, it was Jean-François Lyotard who clearly identified, already in the 1970s, the major challenge that databases, artificial intelligence, and robotics would pose to classical authorship.
Lyotard’s celebrated 1979 book The Postmodern Condition: A Report on Knowledge depicted a scenario we are all too familiar with today: the loss of human authority over the production of knowledge, as computers learn to perform tasks as wide-ranging as copyediting and medical diagnoses.1 No longer is the human the producer and the subject of knowledge. The postmodern episteme was characterized by a disorientation whose uncertainty and insecurity were precisely what the modern subject wanted to excise from systems of knowledge. Indeed, one of Lyotard’s targets is Descartes, who preferred Amsterdam’s systematic grid to childishly disorganized Paris. The Cartesian subject seeks the absolute ground that guarantees clarity and distinctness, while the postmodern subject is no longer even human. In other words, postmodernism, with its anti-humanist provocation, already acknowledged the prospect of a post-human era.2
The first pages of The Postmodern Condition already seem to grasp what would happen in the following fifty years—and also demonstrate that it was a mistake to regard postmodernism as merely the cultural logic of late capitalism, without anything profound to say about technology. It is important to point out that Lyotard’s book was a response to a report written by Alain Minc and Simon Nora for French president Valéry Giscard d’Estaing. Published a year before Lyotard’s acclaimed book, L’informatisation de la société (The Computerization of Society) anticipated today’s discussions on digital sovereignty and national technological strategy.3 Almost fifty years later, we encounter the same scenario, but with much greater intensity and urgency: from civic to military applications, the computerization of society is accelerating rapidly. Technological power is reflected in the robustness and efficiency of infrastructures that demand constant optimization, total mobilization, and maximum resilience. Global competition now hinges on microchips, rare earth minerals, and data centers, which in turn reinvent the Schmittian concept of sovereignty: the sovereign is he who decides the exception for the purpose of self-preservation. The state of exception, which is the manifestation of the sovereign, has been instituted by cutting off flows of migration, knowledge, and technology, by disrupting the protocols and market mechanisms that have characterized globalization since the fall of the Berlin Wall.
The dominant impression is that, under so-called sovereign AI, machines in the near future will outcompete human intelligence and come to reign in almost all domains of life, since they are exempt from making errors caused by sympathy, greed, vanity, and human emotion more generally. This challenge puts into question, firstly, the primacy of universities in the production and dissemination of knowledge. If knowledge production can be automated, and teaching can be done by generative AI that turns out to be more personalized, patient, and informative than many professors, the university could be reduced to a technical department for maintaining machines, while perhaps offering students training in core human tasks on the side. Indeed, large language models can already process knowledge that isn’t perfectly accurate and precise. (With its epistemological limits—known casually as “hallucinations”—AI may not be capable of guaranteeing exactitude any time soon.)
In recent decades, we have read theoretical efforts to decenter the human and introduce less human-centric forms of knowledge, from actor-network theory to multispecies perspectives to anti-correlationism. Today, the academic crowds that were fascinated by anti-humanism and anti-anthropomorphism seem paradoxically confused by the industrial politics of artificial intelligence. In fact, a trap has opened up for an anti-humanism or anti-anthropocentrism that believes that modern social crises are largely caused by humanist and human-centered desires to order everything according to human will. On the surface, the shift in the authorship/authority of knowledge seems like a remedy to the all-too-human aspects of modern Western philosophy. But the clearing created by anti-humanism is too quickly occupied by a transhumanism that sees the human as something that needs to be overcome, and a computationalism that sees it is but a computer among many other machines. Yet the transhumanist and computationalist critique of humanism conceals a latent humanism that longs for control and domination.
Indeed, the current hype around artificial intelligence has revived last century’s discourses on intelligence as computational and life as computable—as if the criticisms of the last century don’t apply. This hasn’t happened without a reason of course. We have to recognize that if computationalism can appropriate anti-humanism and present itself as its heir, it is because the various strands of anti-humanism rarely took technology into consideration, save for a few figures in the history of philosophy who recognized the human being as a fundamentally technical being.4 The human is not an antithesis to technology; rather, the process of hominization entails an organological struggle, namely, the effort to incorporate tools into the operation of the human body. This was what Bergson called “creative evolution,” which understands the elaboration of inorganic nature and the invention of artificial organs as a necessary tendency of the development of vertebrates. This contrasts with an understanding of evolution as an adaptation to the organism’s environment and a consequence of competition. For Bergson, toolmaking belongs to intelligence, which has to be distinguished from instinct. Birds make nests, and they do this not by learning the skill from their parents, but by building from their instincts. The human organological struggle, on the other hand, is a learning process, an educational process.
2. The Conflict of Organs
I use the term “organological struggle” to describe what guides us to reflect on our relation to the technological world, as well as to the life of the spirit. Today we live in an environment covered in sensors and digital devices that capture data and record interactions, forcing us adapt to the functioning of interfaces, database structures, and predefined rules. This pressure to adapt gives rise to a countervailing need for strategies that help us adopt these tools for our own use.
With their potential to resolve personal and planetary problems, recent developments in AI give new life to computationalism and enthrall politicians and strategists, who easily get intoxicated by the prospect of transforming the world. But these developments also gradually shift the paradigm of industrialism in a way that goes largely unnoticed, because they appease the imperatives of market-driven democracy and bolster industrial monopolies that are in service of the sovereign. This results in two views on the relationship between the new wave of industrialization and the future of knowledge:
1) Universities should be reengineered according to AI, with not only daily administration replaced by AI but also pedagogy taken over by more personalized AI assistants.
2) The humanities can be eliminated since they don’t contribute to the economy. Even worse, they represent a resistance against technological acceleration.
The first view targets not only teaching and creative professions. Jensen Huang, the CEO of Nvidia, recently claimed that there is no longer any need to learn coding, since everyone can be a programmer with the aid of AI. What is more important to learn, said Huang, are the physical sciences, to know how “laws of physics, friction, inertia, and cause and effect” work.5 The first view also applies to research; as Chris Anderson already claimed almost two decades ago, “with enough data, the numbers speak for themselves.”6 In the past decade data science has emerged as one of the most popular majors at universities.
The second view might recall the early twentieth-century Italian Futurists, who began as anti-traditionalists and became fascists. But Marx reminds us that history repeats itself—the first time as tragedy, the second time as farce. Not only in academia has it become fashionable to laugh at the Luddite left; this tendency has become generalized, as a hatred for any position critical of technological acceleration. Today, with their aura of anti-humanism, the accelerationists are assured and bold. When everything is considered computable, and speaking critically against AI risks the accusation of mysticism, conditions are ripe for the expansion of the nihilist impulse to forge a new world order by aestheticizing war and destruction. AI hype extends computationalism into an industrial ideology like transhumanism, with the difference being that transhumanism is an extension of consumerism, while computationalism is a contemporary paralogism. Worryingly, the two views enumerated above influence how politicians think about education and workplace management.
The problem is that these views are not completely wrong. The transformation of the university by technology is as inevitable as the technological challenge to the authority of the author and the need for new regulations regarding student exams. But reengineering universities around industrial agendas forgets that contributing to economic growth and social impact are secondary to the fundamental function of the university as a place where knowledge—and more importantly, reason—are contested, and where individuals participate in the life of the spirit through collective learning, debate, and synthesis. Knowledge doesn’t only serve as a means to a practical end, like finding a job (even though many universities consider employment rates as a measurement of graduates’ happiness). Rather, knowledge allows us to realize a higher end, one that demands both contemplation and action. This elevation of the spirit is threatened by nihilism, the depreciation of the lofty values once held by the individual and the collective, as Nietzsche elaborated throughout his work. I would argue that the profound importance of digital technology is entirely different than the apocalyptic reduction of everything to calculation, under the banner of computationalism, transhumanism, or “post-humanism.”
It might be helpful to turn to Kant, especially his late work The Conflict of the Faculties (1798). In this collection of three essays, Kant addressed the conflict between the lower faculty of philosophy and the higher faculties of theology, law, and medicine. Kant examined these latter three disciplines because scholars in these fields, whom he calls “businessmen” (Geschäftsleute), can seem like magicians, luring the masses into believing in dogma or even superstition by undermining their faith in reason (i.e., philosophy). In a rare note of humor, Kant writes:
As for the philosophers’ twaddle, I’ve known that all along. What I want you, as men of learning, to tell me is this: if I’ve been a scoundrel all my life, how can I get an eleventh-hour ticket to heaven? If I’ve broken the law, how can I still win my case? And even if I’ve used and abused my physical powers as I have pleased, how can I stay healthy and live a long time? Surely this is why you have studied—so that you would know more than someone like ourselves (you call us laymen), who can claim nothing more than sound understanding.7
Kant is not saying here that philosophy, to be philosophy, must remain devoid of any utility; it should in fact be useful to the higher faculties, he says. However, its utility to the government is a secondary purpose of philosophy, while such utility is the primary purpose of the higher faculties. Can we see how these issues emerge again today in the rhetoric of transhumanism? The singularity is near, as is immortality—don’t you want to take the AI pill and skip your college classes while still graduating with honors? Don’t you want to be immortal without wasting your effort on eating a good diet and exercising? Don’t you want to solve all your personal and legal problems by delegating them to an AI agent? Kant’s conflict of the faculties points to the modern form of consumerism, even if that term was not available in Kant’s time. With Kant’s help, we might ask how we can confront today’s technological restaging of the conflict of the faculties. The key is faith, or belief—not in AI, but in reason, keeping in mind reason’s skepticism regarding facts. Whereas critical philosophy aims to overcome empiricism, which is normative and vulnerable to contingency, AI is based on probability and the association of facts (data). Instead of questioning the limit of artificial intelligence, people are turning to science fiction for salvation.8
Faith in reason means recognizing the autonomy of reason, which consists firstly in acknowledging its limits—acknowledging the things it can know and the things it cannot. Secondly, faith in reason consists in pursuing moral progress: the highest good and universal happiness. It is by recognizing its own limits that reason distinguishes itself from superstition and the magical conjuring of the higher faculties; and it is by aiming for the highest good that reason transcends the limit of theoretical speculative knowledge. Reason acknowledges the indemonstrable rational ideas (the immortal soul, freedom, or God) that are central to practical moral reason. This acknowledgement only comes from within reason itself, not from outside. Rejecting faith in reason and its ends can only lead to nihilism.
Kant’s faith in reason stands in contrast to recent claims that belief in progress means mobilizing AI towards the goal of Western domination. This view is exemplified by Peter Thiel, cofounder of Palantir, and by his fellow cofounder Alex Karp (who is often associated with Frankfurt School thinkers because he studied at Goethe University in Frankfurt). Proclaiming that the West has lost its belief—in God, in religion, in freedom—Thiel and Karp insist that it is only by restoring true progress, lost since the modern era, and by developing technology to serve the nation-state that the West can be saved from its own decadence. This progress means only domination, a return to the time when Western modernity determined the course of humanity. For Kant, by contrast, progress can only be moral progress, which is also the telos of reason. In this sense, the nationalist and fascist appropriation of a national technological agenda in the name of the West is ultimately anti-Western, if not anti-reason. How then can we understand the relevance of Kant’s reason to the question of automation? To transpose Kant’s analysis to the realm of AI, it is first necessary to understand the limits of AI’s particular form of reasoning, now largely represented by LLMs; and it is then necessary to transcend this limit by making use of AI to pursue higher ends. Just as reason must speculatively or theoretically confront its own limits in order to provide a positive definition of freedom, God, and the immortal soul, AI has to expand in order to progress towards moral ends.
3. The End of Machines
How are we to confront the production of knowledge today, in view of digital acceleration? Are we nostalgic for the Enlightenment and its humanism, and for a belief in the autonomy of the human subject? Not necessarily—but rejecting a certain stereotype of the Enlightenment doesn’t mean subscribing to the Dark Enlightenment.9 The constant upheaval of the post-Covid world seems to have put an end to a dream of the 1990s, demanding a new horizon of thinking. If future knowledge production will indeed be delegated to AI, then this will be a kind of knowledge that doesn’t concern us. Such a machine takeover would be a rational process of evolution that we should not try to compete with any more than we should try to outrun a speeding car. After these developments, however, if we are still talking about knowledge, it will be a knowledge of life—of how to live well and how to live together well. Living is, after all, not opposed to death, but to a loss of meaning—the loss of a higher purpose beyond utilitarian ends.
This conversation about the knowledge of life won’t be possible without a renewed critique of political economy, which is fundamentally an organological struggle underlying the long process of anthropogenesis. The organological struggle is an attempt to appropriate technology (i.e., invention and tool use) to produce negentropy, a form of resistance against the entropic process of cosmic life. Nicholas Georgescu-Roegen criticized neoclassical economic theory, proposing instead a bioeconomy based on thermodynamics, in which human beings would invent and use exosomatic instruments to produce negentropy. Yet industries earn profits from entropic acceleration, leading to the destruction of the environment and the erosion of mental health. The recent court case accusing social media companies of trying to get young people addicted to their products is only the tip of the iceberg, since it is not only young people who get hooked on viral videos. Artificial intelligence becomes artificial stupidity when it is mainly used to predict and determine user behavior, or to market and surveil. There is nothing technical about this form of stupidity. It is a perversion born of a technological fanaticism that reduces life to computation, human will to patterns, knowledge to data. Returning to humanism, however, is not a way out, since this would limit any new discourse on the subject. After the critiques of humanism and anthropocentrism, we have to look for a different path, while avoiding the traps of computationalism and transhumanism.
To find this path, the human will firstly have to recognize itself as a technical being whose life is made possible by prosthesis. Yet not all forms of prosthesis are desirable; some resurface a classical opposition between technology and nature in the human world.10 Secondly, the human will have to develop a new political economy that is adequate to the specificity of generative AI and that can redirect it to a higher purpose. We can call this higher purpose “eudaemonia,” Aristotle’s term for human flourishing, which is akin to what Amartya Sen and Martha Nussbaum call “capability.”11 When I say that we need to recognize the “specificity” of the current form of automation, I mean that we have to understand it as radically different from the seventeenth- and eighteenth-century stereotype of the machine, and also different from the categories that Marx offered us. We have to recognize the actual technical reality made possible by digital communication networks, and understand both the advancements and limits of algorithmic automation.12
Since the teaching of reason is already so rare and difficult today, universities that want to instill it in their students will confront a fierce challenge, above and beyond the rationalization process described by Max Weber.13 We are seeing a repeat of the struggle against positivism and mechanism from the late nineteenth and early twentieth century. The spiritual, however, is not opposed to automation; as Bergson says, the spiritual is always implied by automation. Mysticism in this sense is not superstition but rather an insight that emerges when one sees the limits of the dominant technological tendency—whether it’s the limits of mechanism and positivism in Bergson’s time, or the limits of computationalism and transhumanism in ours.14 As Bergson wrote:
So let us not merely say … that the mystical summons up the mechanical. We must add that the body, now larger, calls for a bigger soul, and that mechanism should mean mysticism. The origins of the process of mechanization are indeed more mystical than we might imagine. Machinery will find its true vocation again, it will render services in proportion to its power, only if mankind, which it has bowed still lower the earth, can succeed, through it, in standing erect and looking heavenward.15
The automation of knowledge, with its slogan “death to the author,” leads to a new conflict of the faculties, which is taking place not only in today’s universities but also in our bodies and throughout society. In contrast to Kant’s time, today this conflict originates from an industrialization powered by a digital optimism. The magic of businessmen is seductive, but the lower faculties have to be defended—not to eliminate the conflict of the faculties but to intensify it, which will ensure that the conflict will contribute to an individuation that is open to the future.
Jean-François Lyotard, The Postmodern Condition: A Report on Knowledge, trans. Geoff Bennington and Brian Massumi (University of Minnesota Press, 1984).
Lyotard expressed these tendencies more sensually in his 1985 Centre Pompidou exhibition “Les Immatériaux.”
Alain Minc and Simon Nora, L’informatisation de la société (La Documentation Française, 1978). Translated as The Computerization of Society (MIT Press, 1980).
See Bernard Stiegler, Technics and Time, vol. 1, The Fault of Epimetheus, trans. Richard Beardsworth and George Collins (Stanford University Press, 1998).
“Nvidia CEO Jensen Huang Would Not Have Studied Computer Science Today If He Were a Student Today (sic). He Urges Mastering the Real World for the Next AI Wave,” Economic Times, July 21, 2025 →.
Chris Anderson, “The End of Theory: The Data Deluge Makes the Scientific Method Obsolete,” Wired, June 23, 2008 →.
Immanuel Kant, The Conflict of the Faculties, trans. Mary J. Gregor, in Kant, Religion and Rational Theology, ed. Allen W. Wood and George di Giovanni (Cambridge University Press, 1996), 253.
For a detailed analysis of Kant’s relevance to AI, see Yuk Hui, Kant Machine: Critical Philosophy after AI (Bloomsbury, 2026).
See James Duesterberg, “Silicon Valley’s Favorite Doomsaying Philosopher,” New Yorker, February 18, 2026 →.
According to Simondon, anthropogenesis takes place through an organogenesis that is made possible by nature—i.e., pre-individual reality, which is cosmic.
Martha Nussbaum, Creating Capabilities: The Human Development Approach (Harvard University Press, 2011). The late Bernard Stiegler wanted to explore this project further.
See Yuk Hui, Recursivity and Contingency (Rowman and Littlefield, 2019), in which I outline a philosophical history of machines, from the mechanisms of the seventeenth century to the cybernetics of the twentieth century and beyond.
Towards the end of the Critique of Pure Reason Kant said that mathematics can be learned, but not philosophy (unless in a historical manner). One can only learn how to “philosophize.” Immanuel Kant, Critique of Pure Reason, trans. Werner S. Pluhar (Hackett, 1996), A837, B865.
For more on Bergson’s theory of mysticism and mechanism, see chapter six in Yuk Hui, Machine and Sovereignty (University of Minnesota Press, 2024).
Henri Bergson, The Two Sources of Morality and Religion, trans. R. Ashley Audra and Cloudesley Brereton, with W. Horsfall Carter (Greenwood Press, 1935), 267.