Why is AI the culmination of Modern Gnosticism?

I consider Eric Voegelin’s “Modern Gnosticism” thesis the crowning point of his work. I know that the thesis is the most controversial part of his work, and even many of his close followers reject it – but, as they say, this is their problem. Voegelin’s visionary insight – as this is not just a mere idea: the entire Order and History project was conceived beyond the modern, neo-Kantian “idea” of an “idea” – reaches the very heart of the modern world, and is the culmination, and in a way solution, of Nietzsche’s concern with nihilism – which soon led him towards madness; and Weber’s concern with the “religious rejections of the world” – which was theorised in an intermediate piece of reflection, termed Zwischenbetrachtung, written 1914-15, and which was not followed up as Weber died in the Spanish flu after WWI in 1920.

Gnosticism, any Gnosticism, is of course about knowledge; about mere or pure knowledge (thus the affinities with the thinking of Kant are immediately evident); about the saving power of such search for mere or pure knowledge: knowledge at any price, any knowledge, or search for it, being justified – based on the background assumption, which sometimes is brought explicitly forward, but is always the hidden basis of the entire search, that the world, as it is, is insufficient, faulty, chaotic, unreliable, erroneous, and thus must be corrected. The difference between modern and classical Gnosticism is that the shortcomings of the world in the ancient version are explicitly named, and the creator god is identified as an evil demiurge, while in the modern version this original narrative is omitted, and the focus is “merely” on the shortcomings of the actual world, or “Nature”. The hidden motto of such a striving is that “we must make the world a better place” – a motto which for some people, in certain circles, is so evident that its questioning is considered to amount to madness, or evil thinking.

AI so clearly fits into this line of conceiving the world that this does not require much discussion. Let me only mention one simple point, self-driving cars. Humans, as we all known, are imperfect creatures. They are capable of all kinds of mistakes behind the wheel: driving too fast, driving under the influence of spirits, getting distracted, even falling asleep. So they cannot be allowed to drive their own car, and should be replaced by AI-driven robots, which – presumably: and this is already a tremendous presumption – are incapable of erring.

This is just an example, and I do not want to get caught in the trivialities concerning the pros and cons about this point.

Rather, I will try to show that the very logic which AI follows is not just limited, but is technically Gnostic.

The first step here is that everything in the world – this point is emphatic and is not taken lightly: EVERYTHING in the world – is reduced to information, or a piece of knowledge, which can be precisely identified and coded. Here the connection between AI and internet information gathering is very tight, as AI is ideally using any information about anything and anybody that can be available through the internet.

This is both much, even too much, and is precious little.

It is too much, as Internet activities contain an enormous amount of trivialities. The situation reminds one of the “demon of second order” in Stanislaw Lem’s Cyberiada, a great book written under Communism, but which has a surprising actuality in our own days. Gaining something meaningful out of such activities also reminds one of the way archaeologists try to discover past cultures and civilisations without writing from the remains – say, the garbage heap – that they left over. But it is therefore also precious little, as everything that really matters, that gives meaning to human life, falls outside the possibility of codification.

This point is so simple and trivial that one feels ashamed that it is necessary to talk about it, for reasonable adults. Adherents of AI should be given to read The Little Prince by Antoine de Saint Exupéry, to test their ability to understand it. Generations of children were brought up, all around the world, by this book, and I challenge any AI to make any sense of it. AI-s similarly cannot, and will never be able to, understand the smile of a child, or the meaning of a particular gesture or facial expression. Dickens is spending hundreds of pages in his novels to introduce characters whose acts, gestures and expressions can then be understood – though not in a manner that any AI would ever be capable of understanding. All the more as AI-s are incapable of any genuine understanding – they can probabilistically identify patterns, and can then extrapolate this in massive quantities, and in this way can make money for those who is using them “professionally”. There is no question about this. But this is not understanding. And it does not lead to anything meaningful.

Let me offer another example, through poetry. Poetry is using language; furthermore, every great poem has an exact text and structure, so it can be perfectly inserted into a machine. Yet, no AI will ever produce great poetry, as all that an AI can do is to imitate whatever has already been done, but cannot become a poet, with its unique, non-confoundable style. And, more than that, it is great that it is so, as if it were possible to artificially produce poetry, life would cease to be meaningful, and we would move from an AI incapable of meaning to an AI destroying meaning. An AI can give answers to an essay question that will result in high grades. But it cannot write poetry. And this is important, as poetry, arguably – and many people argue in this way – is the greatest expression of the human spirit, more so even than philosophy – but great poetry is philosophy itself (see Heidegger on Hölderlin); or, great philosophers are themselves poets – see for e.g. Nietzsche.

Let me turn to a very simple example. Great poets often repeat not just words but phrases; and not just once but twice. Or more times. Examples are so abundant that they are meaningless to evoke. Now, as we all now, through our own computer, if we repeat a word, our computer immediately marks this as an error. But in a great poetry it is not an error (and it isn’t often even in my own humble writings). Which AI would ever be able to notice the meaning of such an evidently “meaningless” repetition? But in poetry it isn’t meaningless because the first repetition has a different meaning then the original statement; and the second repetition has a still different meaning. Human readers realise this – with something that the ancients, and for a good reason, called their “soul”.

Note that Lem’s Cyberiada even foresaw the possibility of computers writing poetry, though only of a very particular kind – and it was part of a farce, not the presentiment of a machine writing genuine poetry.

The striking thing about AI is not the presumed earth-shaking enormity of its discovery – as, if the saying that “there is nothing new under the stars” can be applied to something, this something is AI. AI is simply the perfectioning of statistical and computational methods whose development is going on since decades, and which ultimately is only “helpful” as an economic project – some people can make an enormous amount of money by this; and as an political project – it provides an increasing possibility to control people: this is why Communist China and Western power-centralising agencies progress almost hand in hand in this endeavour. What is striking is how it is being widely accepted, without minimal “critique” or “protest”.

Here two things are important to explore. The first concerns the way in which the acceptance of the AI logic was prepared. As I have already argued, the very mentality underlining AI involves a frontal collision with the core values of European culture – just as, in fact, with most other cultures. The acceptance of the AI logic required a long process of deculturation – which is indeed going on since decades, with universities abandoning programmes in the classics and the humanities, and high schools no longer offering an adequate programme in the teaching of premodern history or classical literature. This centralised effort was much helped by the way critical and avantgarde movements and intellectuals favoured the politics of gender and sexuality, and from that optic promoting recent literature of rather questionable value, instead of a proper education in the classics. Thus, again, the complicity between the left-wing critical avantgarde and the technocratic-financial elites is evident.

Here a crucial role was played by the newly dominant HR Departments, which suddenly became mushrooming everywhere (who would write the history of their rise to dominance?), and which not only promoted the absurd idea of formulating “learning outcomes” (as if professors could, and should, tell in advance what students would learn from their courses), but extirpated the word “understanding” from such descriptions as being “too vague”. This is an extremely telling development, as “vagueness”, on the one hand, was a major charge formulated by neo-Kantians like Heinrich Rickert against Wilhelm Dilthey, leading to the lamentable marginalisation of the latter in German philosophy (though there is no question about Dilthey being the much more important philosopher); on the other hand, it is clearly part of the development towards reducing thinking to exact-like stereotypical formulations which can be easily formalised in computer programmes. Neo-Kantianism, neo-positivism, neoclassical economics and analytical philosophy all prepared intellectually, often explicitly, the path toward statistical computation and AI, marginalised classical European culture in the exact same way as Marxists and other radical avantgarde movements did.[1]

This leads to the second point, which is the undermining of the capacity for meaningful protest. Since centuries, progressive left-wing avantgarde movements appropriated any possibility for expressing dissent, by their endless and mostly improper or meaningless social and political crusades (culminating the in current politics of sexuality). The exception, to some extent, was the East-European dissident movement, which indicates that some unused possibilities can still lie there.[2] But in the West, dissident movements were so emptied by permanent and meaningless protests that now, when a particularly absurd modality of the leftist avantgarde joined forces with the worst aspects of the technological-military-industrial complex, the very possibility of dissenting from such developments seems to have been pre-empted.

So now let me try now to explain as clearly as possible what AI is really doing, and what it exactly means. And let me start by stating (and I apologise in advance for such an explicit presumption) that I claim special competence in doing this. I started my academic career precociously, as an econometrician turned quantitative sociologist, at the age of 21, in a research group at the Institute of Sociology of the Hungarian Academy of Sciences which, if it did not want to save the world, it certainly considered as its mission to bring genuinely scientific Western methods into the intellectual desert of Communist Hungary. However, within about two and a half years, I realised that what we were doing, with the greatest zeal and intellectual seriousness, is all but meaningless, as by collecting “empirical data” and running computer programmes we could never discover anything meaningful about reality – Voegelin’s “rrreality!”,[3] as I discovered much later. What we were doing were just approximations, and this did not satisfy me. We were identifying, forecasting and explaining averages, and what I came to be interested were the outliers – deemed as “scientifically uninteresting”. But was it really so? Were not all great human beings “outliers”, in one way or another? And not only the great figures of intellectual or political history, but all those simple genuine human beings I encountered in my own life? What is the meaning to reduce everything to numerical averages, and probabilistic calculus, as reference points?

The point I was wondering about, or which made me change my academic interests, and my life, was formulated better, with precision, by Simone Weil,[4] about the ‘negligible’, which I encountered much much later through Roberto Calasso. The ‘negligible’ is the residue of any codification, or the ‘outlier’ of empirical quantitative research; whatever falls outside a generalist framework; whatever cannot be explained by the model. Yes, but says Weil, “the negligible is the world”: meaning life, the person, the concrete, whatever gives meaning. And so it is central; it is the centre; it is all that matters.

The statistical-probabilistic thinking behind AI denies life, denies the concrete, denies whatever has happened (thus denies history), and denies the very value of events, encounters, life accidents. AI is therefore technically nihilist, in the sense of Nietzsche, as it rejects the world as it is, claiming to vastly improve it in a universal and practically unlimited scale, so it is comparable to the world rejecting religions, according to Weber – it is therefore a quasi-religion, just as communist was, according to many, including Voegelin. The related idea of a “technological religion” was introduced convincingly in a recent VOEGELINVIEW essay by Paul O’Connor.[5]

It is important to mention here – but I can only evoke the point, cannot analyse it fully – that probabilistic theory was discovered by Pascal, one of the most explicitly Christian philosophers of modern times, connected to his famous wager. Probabilistic thinking for him served to undermine determinism – Pascal was unreservedly hostile to Descartes: he characterised him with the two words “uncertain” and “useless”; and to call attention to the infinite variety of possible events in life and in nature, or the created world. He could not foresee that exactly the line of thinking opened by him would become a way to entrap human lives – and more than the Puritan “iron cage”.

This “super-iron-cage” works by using past “experiences”, anything from the past that can be codified and turned into an information, to predict and in this way to form (perhaps the real meaning of “information”, originally implying something like “shape, give form to”) further possible events and forms of conduct, and turn this into a monetary value or a means of control, thus power. But this means that any future possibilities will be as if pre-empted and enclosed, entrapped, into the prison forecast by past actions and events.

This process can be jointly captured by two seemingly opposite metaphors, liquidation and freezing. “Liquidation” is an extremely important term, connected both to money (as “liquidity”), and destruction or execution (note here also the rise to dominance of the word “executive”, but also note the “liquidation” was a favourite term used in the Communist purges – a kind of euphemism for killing), but also to rendering whatever solid is liquid, much used by Zygmunt Bauman in his influential books. The central aspect of the metaphor is the undermining and destruction of any solidity or stability, thus the destruction of the possibility of meaning (stability is the precondition for meaning: a relationship that is not stable cannot be meaningful; while concerning the word “solidity”, let us not forget the direct connection to solidarity, centrally including Polish Solidarnosc, a movement directly inspired by Pope John Paul II). Under conditions of liquidity, it is impossible to establish meaningful relations, so one is reduced to make money or gain sensual pleasures, the two driving forces of hypermodernity (and the presumed unique driving forces of the human being, according to the dominant ideologies of hypermodernity). Freezing, on the other hand, implies the end of any movement, an imprisoning into unalterable and lifeless structures. The two terms in facts go together, forming a process, of which the first point is the liquification of all stable structures and relations, animated by a presumed drive to freedom (interpreted as a jump into the void), which then can be safely and forever frozen – by the statistical-probabilistic models on which AI operates. The end point, and the end of life, is reached with the absolute freezing point: not the “end of history”, but the outcome of a reliance on AI. Where our current technological, scientific, economic and political “avantgarde” is leading us – as cattle to the slaughterhouse.

So, I think we need to go a step further, in line with the underlying thrust of the Modern Gnosticism thesis, beyond quasi-religions. Classical Gnosticism, in its limitless hatred of the world, wanted either to escape it or destroy it. And it is here that the core of AI as the culmination of Modern Gnosticism lies. I do not know whether there are forces in the background that coordinated the rise to dominance of HR Departments, the politics of gender and sexuality, and AI. But it is clear to me that the promotion of AI involves world destruction – perhaps even as a project.

Life, and Nature, to repeat, is unique, in its concreteness. It cannot be improved, in the sense of replacement or substitution, only bettered in the sense of ennobling. Accidents should be mended, but they cannot be fully prevented, and the entire mentality, and mania, of prevention, or “prevent-mentality”, is itself life-hostile and nihilistic, based on the universal application of the same probabilistic logic which annihilates concreteness and meaning.

Thus, if ever, we are now facing a Kierkegaardian situation: either-or. We either let AI rule us and take away our lives, destroying our world – or we realise that the entire line of development culminating in the rise to dominance of AI is mistaken, a kind of entrapment in a trickster logic, part of Modern Gnosticism, and thus should be limited or – better – completely abandoned.

Turn around. Periagoge.

Notes

[1] For the crucial role played here by William Stanley Jevons (1835-1882), founder of neoclassical economics and modern logic, through the generalisation of the idea of “substitution”, just as the builder of the first computer, see Arpad Szakolczai, ‘Substitution: The animating spirit of economic rationality’, in Theory, Culture and Society 42 (2025), 3: 3-20; doi: 10.1177/02632764241278887 (open access).

[2] Jan Patočka, Pope John Paul II, and Sándor Márai are three particularly important figures in this direction.

[3] See the anecdote told by Gregor Sebba, in his “Prelude and Variations on the Theme of Eric Voegelin,” in Ellis Sandoz (ed.), Eric Voegelin’s Thought: A Critical Appraisal, Duke University Press, 1982, p. 89.

[4]Simone Weil was a rather perplexing writer, and person, but she had a major impact on two of the most important thinkers of the past half century, Michel Serres and Roberto Calasso, whose work is way outside and beyond the dominant contemporary approaches of analytical philosophy and critical theory. For recent, excellent work on Weil, see Wolfram Eilenberger, The Visionaries, Penguin Press, 2023; and Harald Wydra, “Defending the Needs of the Soul: Simone Weil ‘s Anthropology of Grace”, International Political Anthropology 19 (2026), 2, forthcoming.

[5] See Paul O’Connor, “Technological Religions”, VoegelinView, June 1, 2026.

Print